US20120331561A1 - Method of and Systems for Privacy Preserving Mobile Demographic Measurement of Individuals, Groups and Locations Over Time and Space - Google Patents

Abstract

Disclosed are techniques for privacy preserving mobile demographic measurement of individuals, groups, and locations over time and space. A method of estimating demographic information associated with a user of a mobile device and/or a location while preserving the privacy of the user based at least in part on a location estimate of the mobile device of the user includes receiving an estimated geographical location of the mobile device of the user and receiving a time at which the mobile device was at the estimated geographical location. The method includes assigning substitute identifiers for the geographical location and the time at which the mobile device was at the estimated geographical location. The method includes associating the geographical areas substitute identifiers with demographic information and estimating demographic information associated with the user of the mobile device based on the substitute identifiers and based on the demographic information associated with substitute identifiers.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
• This application claims the benefit under 35 U.S.C. §119(e) of U.S. Provisional Patent Application Ser. No. 61/499,975, filed on Jun. 22, 2011, entitled Method of and Systems for Privacy Preserving Mobile Demographic Measurement of Individuals, Groups and Locations Over Time and Space, herein incorporated by reference in its entirety.
• This application is related to U.S. patent application Ser. No. 13/252,685, entitled Method of and System for Estimating Temporal Demographics of Mobile Users, filed Oct. 4, 2011, Attorney Docket Number 2000319.174 US2 which is incorporated by reference herein.
BACKGROUND OF THE INVENTION
• 1. Field of Invention
• The present disclosure is in the field of demographic, psychographic, and behavioral profiling of individuals and locations based on mobile device movement. The present disclosure includes methods of obtaining both individual and location profiles while preserving the privacy of the individuals.
• 2. Description of Related Art
• Advertisers, marketers, and businesses attempt to match their products with the most likely buyers of the product. In order to do this, they use information such as the age, the buying power, the activities, and many other demographic, psychographic, and behavioral information on an individual or a group of individuals to ensure the best target audience for their product.
• Traditionally in the US, demographic information has been provided by the US Census. This information attempts to measure many attributes of small geographical areas based on household data. This provides a base level of demographic information related to the persons who live in a given area. Over the years, various methods and information sources have been proposed and used to enhance the accuracy and specificity with which this information can be applied to individuals and groups.
• Web sites and online web usage tracking has added a new dimension to the toolbox for these would-be marketers. By utilizing techniques such as browser ‘cookies’, the profile of an online user can be augmented based on their online behavior and the sites that they visit.
• As more and more people make use of smart phones and other mobile devices equipped with the capability to determine location, yet another set of identifiable information can be added to the mix. In particular, the time and location of the device itself can be used to estimate many characteristics of that device's user. By determining the demographics of the locations, venues, and times at which these destinations are visited, it is possible to build more detailed estimations of the individual's demographic, psychographic and behavioral characteristics (See US 2002/0111172 A1, DeWolf, et al).
• At the same time, by analyzing the various device profiles at a given place and time, the class or classes of people that are at a location can be computed. This becomes a demographic profile of that place at a given time.
• For the purposes of this disclosure, the term “demographic profile” refers to a set of attributes describing the user of a device (the device demographic profile, or DDP) or the group of people that visit a particular location (the location demographic profile, or LDP). This set of attributes may include, but is not limited to, age, gender, ethnic background, income, years of education, as well as behavioral descriptors, such as “frequent traveler” or “retail shopper”.
• Demographic profile data for where people live (the LDP) has been widely available for decades. Direct marketing firms and others have compiled data from public and private sources to build profiles of neighborhoods. Public data sources include the U.S. Census, public record filings about home purchases, and records on public infrastructure such as water systems. Private sources include shipping and purchase records, magazine and newspaper subscriptions, and voluntary surveys.
• LDPs are available for nearly every residential block in the U.S. Some vendors of demographic profile information further categorize neighborhoods according to the mix of different behavioral types based on the mix of these attributes—labeling combinations of attributes with names like “urban achievers.” Together these data sources have given marketers a way to target households for various offers by direct mail and telephone, providing vendors an easy way to send their catalogs only to highly educated parents under 45 who live in the suburbs, for example.
• While much is known about the LDP of a block of homes, which can be characterized as the static or unchanging demographic profile of a place, it has been discovered that little is known about where the people living there go during the day, which can be characterized as the dynamic or changing LDP. Some companies have surveyed users to determine this information, and extrapolated from these limited samples. This can give some broad general understanding of where some people work during the day, but cannot resolve how the LDP changes during the day, or over the course of a week, for example. Commercial and urban areas in particular, where the demographic mix changes significantly with time of day or day of week, cannot be described by current demographic profile data sources. Companies evaluating these areas for retail expansion or outdoor advertising have very little information on which to base large financial commitments.
• User demographic profiles, not associated with one's home, are frequently collected using purchasing and shipping records, web browsing histories, and other records related to computer usage. An example of this is the user profile data collected in web browser “cookies”, which are special files many web sites use to store information about users based on pages they have visited. These files are stored on the user's computer and submitted to the site with every page request or view. Online user demographic profiles are generally computed without considering their location or their current activity.
BRIEF SUMMARY OF THE INVENTION
• In one aspect, the invention features methods of and systems for privacy preserving mobile demographic measurement of individuals, groups and locations over time and space.
• In another aspect of the invention, a method of estimating demographic information associated with a user of a mobile device while preserving the privacy of the user based at least in part on a location estimate of the mobile device of the user includes receiving an estimated geographical location of the mobile device of the user and receiving a time at which the mobile device was at the estimated geographical location. The method also includes providing a set of substitute identifiers for a corresponding set of at least one geographical area, assigning one of the set of substitute identifiers for the geographical area corresponding to the geographical location of the mobile device, and assigning a substitute identifier for the time at which the mobile device was at the estimated geographical location. The method further includes providing an association between the substitute identifiers for geographical areas and demographic information corresponding to the substituted geographical area and estimating demographic information associated with the user of the mobile device based on the assigned substitute identifiers and based on the demographic information associated with the provided set of substitute identifiers.
• In a further aspect of the invention, the method further includes estimating the received geographical location associated with the mobile device of the user.
• In yet another aspect of the invention, the method includes recording the estimated demographic information associated with the user of the mobile device in a device demographic information log. Optionally, the method also includes estimating demographic information associated with at least one geographical area of the set of geographical areas based on the demographic information recorded in the device demographic information log. The device demographic information log contains a plurality of records of estimated demographic information associated with a plurality of mobile devices.
• In another aspect of the invention, the method includes sending the estimated demographic information to the mobile device.
• In still a further aspect of the invention, the substitute identifier for the geographical area corresponding to the geographical location of the mobile device that is assigned identifies a particular set of demographic information.
• In an aspect of the invention, the substitute identifier for the geographical area corresponding to the geographical location of the mobile device that is assigned is reduced in specificity relative to the estimated geographical location of the mobile device.
• In another aspect of the invention, the substitute identifier for the time at which the mobile device was at the estimated geographic location is a measure of time that is reduced in specificity relative to the time that was received. Optionally, the substitute identifier for the time at which the mobile device was at the estimated geographic location is a representation of time lacking date information, a time range, and/or an hour of a week designation.
• In still another aspect of the invention, the estimating demographic information associated with the user of the mobile device is performed on a separate computer system from a computer system performing any one or a subset of the other steps. Optionally, the separate computer system, relative to the computer system performing any one or a subset of the other steps, is maintained in a separate network, maintained in a separate building, and/or maintained by a separate operational entity.
• In another aspect of the invention, a method of estimating demographic information associated with a geographical area and a time period based on demographic information associated with users of mobile devices within the geographical area includes providing a set of geographical areas and providing a set of time periods. The method also includes receiving an estimated geographical location of a mobile device of the user, receiving a mobile device identifier that is associated with the mobile device, and receiving a time at which the mobile device was at the estimated geographical location. The method further includes determining the geographical area of the set in which the estimated geographical location occurs, determining the time period of the set in which the time at which the mobile device was at the estimated geographical location occurs, and retrieving information representative of demographic information associated with the user of the mobile device based on the mobile device identifier. The method also estimates demographic information associated with the determined geographic area during the determined time period based on the retrieved information representative of demographic information associated with the user of the mobile device.
• In a further aspect of the invention, the method also includes estimating the received geographical location of the mobile device of the user.
• In still another aspect of the invention, the method also includes providing a set of initial demographic information associated with the determined geographical area. The estimating demographic information associated with the determined geographic area is further based on the initial demographic information. Optionally, the method also includes, subsequent to the estimating demographic information associated with the determined geographical area, adjusting the initial demographic information based on the estimated demographic information. Also optionally, the initial demographic information is based on governmental census information, public record information, shipping and purchase records, magazine and newspaper subscriptions, voluntary surveys, and/or records of social media activity.
• In yet another aspect of the invention, the method also includes sending the estimated demographic information to the mobile device.
• In another aspect of the invention, the geographical areas of the set are reduced in specificity relative to the estimated geographical location of the mobile device.
• In a further aspect of the invention, the time periods of the set are reduced in specificity relative to the time that was received. Optionally, the time periods of the set are a representation of time lacking date information, a time range, and/or an hour of a week designation.
• In still another aspect of the invention, the estimating demographic information associated with the determined geographic area during the determined time period is performed on a separate computer system from a computer system performing any one or a subset of the other steps. Optionally, the separate computer system, relative to the computer system performing any one or a subset of the other steps is maintained in a separate network, maintained in a separate building, and/or maintained by a separate operational entity.
• In another aspect of the invention, the method also includes performing selected steps a plurality of times for different mobile devices of different users and recording sets of information for the determined geographical areas, determined time periods, and retrieved information representative of demographics information associated with the users of the mobile devices in a device demographic information log. The estimating demographic information associated with the determined geographical area during the determined time period is further based on a plurality of the sets of information in the device demographic information log.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS
• FIG. 1 shows a general flow chart of the process of embodiments of the present invention;
• FIG. 2 shows one embodiment of a deployment of the present invention;
• FIG. 3 shows one embodiment of a detailed example of demographic profile retrieval as deployed on a mobile computing device as perFIG. 2;
• FIG. 4 shows a detailed view of one implementation of data storage on the server for embodiments of the present invention;
• FIG. 5 shows a flow chart of the process of retrieving the demographic attributes of a given location for a specific time;
• FIG. 6 shows a flow chart of one embodiment of the process to compute the Location Demographic Profiles;
• FIG. 7 shows a flow chart of one embodiment of the process to compute the Device Demographic Profiles as deployed according toFIG. 2;
• FIG. 8 shows a second embodiment of a deployment of the present invention in which the Mobile Computing Device requires little to no modification;
• FIG. 9 shows a second embodiment of a detailed example of demographic profile retrieval as deployed on a remote location server as perFIG. 8;
• FIG. 10 shows a second embodiment of a flow chart of the process to compute the Device Demographic Profiles as deployed according toFIG. 8;
• FIGS. 11A and 11B depict an illustrative embodiment of the invention showing aspects of the privacy preserving mechanisms,FIG. 13A represents the Demographic Public System andFIG. 13B represents the Demographic Private System.
• FIG. 12 shows an embodiment of the process to convert demographic data into privacy preserving components;
• FIGS. 13A and 13B depict a modified embodiment ofFIG. 11 showing additional detail with respect to the data elements used for privacy preserving purposes.FIG. 13A represents the Demographic Public System andFIG. 13B represents the Demographic Private System.
• FIG. 14 depicts an embodiment of a partial output of the Location Demographic Profile for an illustrative geographic geometry.
DETAILED DESCRIPTION OF THE EMBODIMENTS
• As used herein, the term “demographics” means statistical data describing a population. Demographics or demographic data includes, but is not limited to average age, income, education, ethnicity, gender, disabilities, mobility, educational attainment, home ownership, employment status, etc. It also may include psychographic data such as the values, attitudes, moods and interests of a population. It also may include activity, venue, and points of interest details that further describe a population or a location. The data may be time dependent with trending attributes.
• As used herein, the term “demographic attribute” indicates a specific measure of demographics, for example, gender. Each attribute may have many “buckets” which represent the possible categories for a given attribute. For example, the attribute gender could have two buckets, male and female. Each of these buckets could indicate the ratio or percent of the population that fall within the definitions for each bucket.
• As used herein, the term “Location Demographic Profile” or “LDP” means the composite or aggregate statistics of a population based on a specific location or geographical area.
• As used herein, the term “Device Demographic Profile” or “DDP” means the composite or aggregate statistics of a Device and, by proxy the device owner or operator(s), based on a set of inputs including the aggregate set of LDPs in which the device has been observed.
• As used herein, the terms “Canonical Week” and “Hour of the Week” refer to the 168 hours of a week starting with hour 0 atSunday 12 AM and continuing through Saturday 11 PM.
• Mobile phones equipped with GPS and other location technologies provide a new opportunity for discovering user information. Location trace information can be used to augment observations about what web pages a user visits or what mobile applications a user employs with respect to time and place. Behavioral and demographic attributes of a user can be inferred based on where he or she goes if a complete map trace is stored. Venues and Points of Interest (POIs) and their demographic tendencies can also be used to contribute to understanding an individual's proclivities. In addition, geotagged social activity (for example, Twitter posts) can provide more social, demographic, activity, and behavior information about a place in time.
• Individual profiling can be used to present relevant advertising and content, both physical and electronic, to individuals. Location profiling can be used to make intelligent decisions about physical places by knowing what kind of people visit a place throughout the day, for example decisions about the best place for retail expansion or outdoor advertising. Such profiling, however, can raise general privacy concerns for the individual, since the location of the individual can tell a lot about the person.
• Tracking of individuals, even without knowledge of their Personally Identifiable Information (e.g. Name, Social Security Number, Address, etc.), presents potential risks to the privacy of individuals. In particular, location information about individuals in which time sequence is maintained and which has high enough spatial precision, can uniquely identify an individual. Linking these unique traces to a specific named person is quite achievable in today's environment of “Big Data”, data that is easily accessed and searched.
• Individual location traces provide enough information to identify the home and work place of individuals. It has been shown in Golle & Partridge,On the anonymity of home, work location pair(Golle, P.; Partridge, K. On the anonymity of home, work location pairs. Proceedings of the 7th International Conference on Pervasive Computing; 2009 May 11-14; Nara, Japan. Berlin: Springer; 2009; LNCS 5538: 390-397), that having this information, even at the spatial resolution of a Census Block, uniquely identifies individuals.
• In order to preserve user privacy, a vendor may choose not to collect any location trace information or associate a unique identifier with any individual locations. However, this makes it very difficult to associate any kind of description or category with a device, since there is no history whatsoever stored about it.
• Embodiments of the present invention disclose methods of and systems for aggregating location and demographic information of individual device users and locations, thereby enabling the various use cases associated with this information, while also preserving the privacy of the individual. Embodiments of the present invention provide a set of methods to preserve the privacy of individuals while computing a Device Demographic Profile based on the Location Demographic Profiles of locations visited by that device. In particular the methods disclosed provide techniques for computing the DDP without storing a location trace of an individual.
• Further, dynamic LDPs can then be determined using a collection of DDPs that visit a location at a particular time of day. These dynamic LDPs can be computed for various scales of time and space.
• In broad terms, embodiments of the present invention include a method and systems to build and maintain demographic estimates of mobile device owners while preserving the anonymity and privacy of the individual. This information is then used to establish dynamic or time-varying demographic information related to a location. Through the use of mobile computing devices in concert with location services, device and location demographic profiles can be computed in a time varying manner without compromising individual privacy.
• Embodiments of the invention determine a Device Demographic Profile based on where that device goes—the interaction with Location Demographic Profiles and potentially other DDPs the device comes near physically—without storing locations, or any trace that would allow discovery of where the device actually went before or after the fact. The only thing stored is the set of demographic attributes for the device (the DDP) and for the location (LDP). No latitude or longitude or description of a location that could uniquely identify a particular point on a map is stored in concert with a unique ID.
• In the first embodiment, the DDP is determined based on the location a device visits at certain times of day. For example, when a phone is used to query location in the evening or at night, for the purpose of looking at a map, or posting a message to a social network, or any other purpose, a simple technique could assume that this is the home demographic of the user, for example. (More sophisticated technique can be created which use statistical measures of time and space to determine demographic clusters and reason as to which of these clusters represent the home demographic profile.) The base Location Demographic Profile for this “home zone” or a set of LDPs chosen for other significance can then be combined to create a DDP. Other LDPs the device encounters can be used to adjust the DDP over time. Embodiments of the present invention determine a DDP without a stored location history thus preserving user anonymity.
• The present illustrative embodiment uses at least three inputs plus an optional additional input. These include the current location as a latitude and longitude point with error probability. The second necessary input includes the available LDPs for the currently or recently visited locations (which may be based on data from the U.S. Census, for example). Third, the current local time, where local is defined relative to the current location. Lastly, the existing DDP as previously computed is an optional input.
• LDPs are initially populated with currently available, static or non-moving place-based demographic data, for example demographic data available from the U.S. Census, but also from other data sources. They may also be extrapolated from neighboring LDPs if not enough data is available for a particular location from other sources. An extrapolation between two areas can be performed in much the same way that a linear interpolation can be performed between two points on a line, taking into account the 2-dimensional nature of the area. The present embodiment adjusts LDPs over time based on the DDPs that visit that location. A feedback mechanism is established that adjusts DDPs and LDPs over time, improving their quality. Individual LDPs are stored for each time of day at a particular place. Additional distinct DDPs can be created and stored which represent various classes or periods of time. This could include such classifications as weekends, morning, lunchtime, afternoon, happy hour, evening, witching hour, etc., and can be defined by any arbitrary block or blocks of time.
• The methods that contribute to the privacy preserving nature of the present invention include A) reducing the specificity of information, B) converting from spatial dimensions to demographic dimensions, C) separating operational domains, D) implementing cryptographic isolation between operational domains, and E) applying probabilistic techniques to further isolate information. These methods contribute to the privacy preserving nature while maintaining the ability to compute DDPs and LDPs while providing the facility to continually add new sources of base demographic or ‘location’ information.
• In one embodiment, the specificity of certain information is reduced. This information includes exchanging the hour of the week in place of the specific local time. This reduces the ability to time sequence a set of observations. This substitution is carried out throughout the entirety of the process.
• In certain embodiments the conversion from spatial dimensions (e.g. latitude and longitude) to demographic dimensions (e.g. age, gender, etc.) occurs at the initial observation. Converting the information early in the process allows the location to never be stored. In other embodiments, the conversion occurs at a later time and before storage of the information.
• In one embodiment, the system is comprised of at least two operational domains. The separation of the operational control provides the ability to reduce information knowledge within each domain. Thus, information that, for example, could link a device with a hashed Device ID is only known in one domain, while information that could link a hashed Device ID to a demographic Profile is only known in the other domain.
• Additionally, the separation of operational domains allows the use of cryptographic techniques to secure private information within one domain from being shared with the other domain. In one embodiment this can be used to create opaque strings which contain very little information for any entity not in possession of the key.
• The use of probabilistic techniques provides another method to increase the privacy of individuals. In one embodiment, probabilistic techniques allow the identification of uniqueness without the ability to determine which of the total set is observed. This technique can be applied, for example, in counting unique devices that appear within a specific spatial area in a specific amount of time without compromising the privacy of an individual.
• Referring now to embodiments of the invention in more detail,FIG. 1 provides an illustrative example of a dynamicdemographic information flow100. It depicts the overall flow of processing and data flow for an embodiment. A software application (generally running on a mobile computing device) requests the location of thedevice101. This request triggers the retrieval of the current Device Demographic Profile (DDP)103. It then computes the location of the device through any of a number of means (e.g. GPS or Wi-Fi positioning)102. This information including the timestamp, latitude, longitude and DDP, is logged to the Device Demographic Profile Logs111 (step104). In order to ensure privacy of the device and subsequently to the end user of the device, no device identifying information is stored in111.
• Continuing inFIG. 1, the current Location Demographic Profile (LDP) is retrieved105 from the LocationDemographic Profiles database112. This information is then logged in the Location Demographic Profile Log113 (step106). Information stored in this database includes the timestamp of the request, the location latitude and longitude as well as the current LDP for that location. The location and optionally the LDP and DDP are returned to the calling application by107.
• Also depicted inFIG. 1 are two processes which do not have to execute in-line or synchronously with the control flows withinFIG. 1. These include the computation of the newDevice Demographic Profiles109 and the computation of the newLocation Demographic Profiles108. These will be described further in subsequent figures.
• In addition, the above steps can be initiated by an express request for demographic information determination by an application (step120).
• Continuing to describe embodiments of the invention in more detail, inFIG. 2 is oneembodiment200 of the invention in which there is aMobile Computing Device201, which containsApplications202. Many applications on mobile devices require location information to be useful or to enhance the functionality of the application. Mobile devices provide aLocation API203, which allows applications to retrieve an estimate of the current location of the device, which, in some embodiments uses aClient Location Service204 on the device. Embodiments of the present invention extend the capability of a mobile device by adding LocationDemographic Profile Logging205, DeviceDemographics Profile Retrieval206 and DeviceDemographics Profile Engine207. Additional functionality are included on a server are grouped within theDemographic Server220. Components of220 include the Device Demographics ProfileLogging221, the LocationDemographic Profile Retrieval222 and the LocationDemographics Profile Engine223. The new components will be described in more detail in subsequent figures. Additionally, the invention, in one embodiment, extends existingLocation Servers210 by augmenting theServer Location Service211 to make a call to theDemographic Server220.
• FIG. 3 is one embodiment of the DeviceDemographic Profile Retrieval206. This figure depicts the process by which a Device Demographic Profile (DDP) is retrieved from the DeviceDemographic Profiles database304. This process is activated, e.g., when a location request is made by an application (step101 ofFIG. 1). Referring toFIG. 3, therequest301 is received and a query is performed to determine if a qualified DDP is available302. The qualification criteria are dependent on several factors including the number of samples that have been collected, the cluster of these samples over time and similar Location Demographic Profiles, and the stability of the sample set with respect to change over time. For example, if a particular sample set shows a rate of change greater than a particular threshold or measure, it can be deemed to be too dynamic to be relied upon. If a qualified DDP is available, it is returned (steps303 and306). In the event that a qualified DDP is not available, a null DDP will be returned305.
• Continuing withFIG. 4, an embodiment of the process by which information is gathered about the demographics of devices (and by proxy the end user of such devices) that visit different locations is described. In particular, each time a location request is made from a device with a qualified DDP, the DDP along with a timestamp or time substitute (such as the hour of the week) and the geographical location or a location substitute (such as a geographic geometry) of the device is captured in the Device Demographic Profile Logs111 by the Device Demographics ProfileLogging221 process. To preserve device anonymity, no device or personal identifying information is logged thus eliminating the ability to track or even identify which device made any particular request.
• FIG. 5 describes theprocess222 of retrieving a specific Location Demographic Profile (LDP). This information is retrieved in order for the techniques to build a history of demographic information about where the device has been observed. A request to retrieve an LDP is received501. A query is performed502 to determine if an appropriate LDP is available for the location of the mobile computing device. The LDP may represent a generalized perspective of the demographic makeup of the location or, for example, it may be specific to the hour of the day and the day of the week. If an appropriate LDP is available, it is returned504, if an appropriate LDP is not present (for example if no devices with known LDPs have visited the location), one may be extrapolated from neighboring LDPs, or a null LDP may be returned by503.
• FIG. 6 represents a process performed according to embodiments of the present invention. The LocationDemographic Profile Engine223 process flow is depicted and describes a method by which various data sources can be combined to create a set of Location Demographic Profiles (LDPs). Each location can have multiple LDPs based on, for example, each hour in the canonical week, the time of day, the day of week, the day of the year, or the season of the year. These demographics are compiled and computed for each location by the depicted process.
• Continuing to refer toFIG. 6, a process to determine the LDPs of one or more locations is initiated. This initiation can happen in real-time based on some event or it can occur in an off-line mode, asynchronous to the operation of the general flow of the mobile device location and demographic requests. For each location that requires a new determination, the location geometry is retrieved from the Location Geometry database610 (step601). The location geometry allows the process to retrieve all relevant information related to that location (where location geometry is simply a portioning of areas in the overall space). This information can include many different and varied data from disparate sources such as Demographic Profile Logs111 from the present embodiment,Static Demographic Statistics605 which may come from sources such as the US Census,Other Psychographic Statistics606 such as interests, activities and opinions of the population that frequents the location,Land Use Data607 which can help identify the types of activity present at the location as well as the expected influx/efflux of people to the location, Venue and Check-indata608 which can provide insight into both the type and frequency of activities in the location, Social Media, News, Crime statistics, etc.609. These alternate sources can provide meaningful measures of activity, inclinations, safety and many other elements that may influence the LDP of the location.
• Continuing to refer toFIG. 6, once the various records related to the location L are gathered (step602), this information is then processed to compute the new LDPs for thelocation108. One embodiment of the technique combines this information using the following technique.
• First, given sufficient Device Demographic Profile samples within the location, for each hour of each day of the week (e.g. Monday at 8 AM), compute the Mobile Demographic Profile for location L by averaging the DDPs within each of the 168 weekly hours.
• Next, determine other hourly statistics (e.g. for social media, these values may change over time and will need to be recomputed).
• Next, combine the Mobile Demographic Profile information with the hourly statistics which represent the dynamic demographics.
• Next, summarize all other static attributes (e.g. Census Data) as the static demographics.
• Next, determine R, the mobile to static ratio for the location L. For example, this can be determined by combining Venue (608), Land Use (607), and Demographic (605) data to estimate the ratio of residents to visitors at the location at the a given time.
• Next, use the ratio R to determine the influence of the dynamic demographics D on the static demographics S. For the set of attribute A that are mutual in these sets, A=(R*S+1/R*D). For those attributes that are exclusive to any of sets, use the values of these attributes as the final values for the LDPs of the location L.
• To complete the discussion ofFIG. 6, once the new LDPs have been computed for the location L, they are used to update the Location Demographic Profiles database112 (step603).
• The process is repeated while there are additional locations L that need to be computed (step604).
• FIG. 7 depicts a process by which the Device Demographic Profiles (DDPs) are computed by the DeviceDemographic Profile Engine207. In one embodiment of this invention, the computation of the DDPs is executed on theMobile Computing Device201. InFIG. 7, an event such as a new LDP or an external trigger begins the process of computing a new DDP for the device. The process begins by gathering all relevant LDPs that have been logged for thedevice701. These may be filtered by different criteria, such as age of reading, day of the week, etc. The logs for the device are then combined using a technique to compute thenew DDP109. One embodiment of the technique chooses all LDPs for the device that were registered between the hours of 10 PM and 5 AM local time for that device. A rudimentary implementation could then average the demographic information contained in all of the LDPs to compute an average DDP. This would presume that the device was near the home location of the owner a majority of the time that it was observed during this time and thus would be strongly influenced by the demographics of that location.
• Another embodiment of the technique would cluster similar demographics based on similarity measures and grouped by time. Thus, large clusters could be identified that represent a demographic profile of locations in which the device is often observed. These different clusters could then be categorized based on various factors (e.g. time of day, land use, etc.) to label these DDPs with tags such as ‘home’ or ‘work’.
• The clustering techniques are novel in that they cluster demographics within the demographic dimension and time dimension. This differs from methods which relied on the spatial attributes (e.g. location) to provide clustering and distance measures. Computations must include a mechanism to determine ‘Demographic Distance’ between any two Demographic Profiles. For example, the distance measure could be the normalized sum of differences in ratios across all buckets in all of the attribute classes. A more sophisticated embodiment computes weighted difference based on the number of buckets contained in each attribute class.
• Completing the discussion ofFIG. 7, once the new DDPs have been computed by109, this information is stored in the Device Demographic Profiles database110 (step702).
• FIG. 8,FIG. 9, andFIG. 10 depict an alternate embodiment of the invention in which the Device Demographic Profiles (DDPs) are computed and stored on theLocation Server210 rather than theMobile Computing Device201. In these instancesFIG. 8 is a modified version ofFIG. 2,FIG. 9,900 is a modified version ofFIG. 3, andFIG. 10 is a modified version ofFIG. 7. One key distinction between these sets of figures is the inclusion of a hashed ID in the process and data flows that identify a device as distinct from all other devices.
• FIG. 8 illustrates an embodiment of the invention in which there is aMobile Computing Device201 which containsApplications202. Many applications on mobile devices require location information to be useful or to enhance the functionality of the application. Mobile devices provide aLocation API203 which allows applications to retrieve an estimate of the current location of the device. Additionally the invention, in one embodiment, extendsDemographic Server220 by adding the LocationDemographics Profile Logging205, the DeviceDemographics Profile Retrieval806 and DeviceDemographics Profile Engine807.
• Continuing withFIG. 9, the modified DeviceDemographic Profile Retrieval806 is described in a embodiment deployed on theDemographic Server220. In addition to the request for a DDP, the calling process must include a hashed ID to identify the particular Device that a DDP is being requested for901. This same information is subsequently required to retrieve theDDP903. All other steps remain analogous as those described for DeviceDemographic Profile Retrieval206.
• Continuing withFIG. 10,1000 depicts the DeviceDemographic Profile Engine807 is depicted in a second embodiment whereby the process is located on theDemographic Server220 rather than theMobile Computing Device201. Similar toFIG. 9, the processing required to compute a DDP on the server is a hashed ID which identifies the particular device for which to compute the DDP. This hashed ID will also be used to retrieve the logged LDPs for thatdevice1001 as well as to update theDDP1002 upon completion of the computation as performed bystep109 previously described.
• FIG. 11A andFIG. 11B combine to depict an illustrative embodiment of the invention highlighting aspects of the privacy preserving mechanisms. This simplified model shows the data and control flow of one embodiment of the system, highlighting those components which contribute to the privacy preserving capabilities.FIG. 11A represents activity occurring on the Demographic Public System1150A.FIG. 11B represents activity occurring on the Demographic Private System1160B. In one embodiment, all actions and data occur on systems independent ofMobile Device1101A. In particular, theMobile Device1101A requests either a Location or Device Demographic Profile or a combination of both from a Location and Demographic Service. TheMobile Device1101A passes either the location of the device or enough information for the service to compute its location in conjunction with a DeviceID that uniquely identifies the device.
• The Demographic Service receives the request and routes it to the Demographic Public System1150A which determines if location information is present1102A. If location information is present, the request is routed to the Demographic Service1104A, otherwise the request is routed to the Location Service1103A. If the location is not present, the Location Service1103A will attempt to compute the location of the device. If it cannot determine the location of the device, an empty (or equivalent) response is returned to theMobile Device1101A. Otherwise, the request is forwarded to the Demographic Service1104A along with the newly computed location. Henceforth, all activities continue as if the request had come from theMobile Device1101A with location included.
• Continuing withFIG. 11A, the information is passed to three independent processes: the Device Demographic Logger1105A, the Location Demographic Counter1114A, and theLocation Demographic Logger1118A.
• The Device Demographic Logger1105A provides a number of privacy preserving actions. It converts the DeviceID into a DeviceToken. It also converts the time of request to the local time based on the location of the device. This local time, in certain embodiments, may then be further reduced in specificity, for example by reducing it to indicate only hour of the week information. Other options include a representation of time lacking date information (such as month, day, and/or year) and/or a time range. The Device Demographic Logger1105A then retrieves the demographic keys for the given location from the Location Demographic Keys1106A. These keys, the converted request time, and the DeviceToken are then passed over anAdministrative Boundary1127 to the Device Demographic Loader1107B inFIG. 11B.
• TheAdministrative Boundary1127 provides additional privacy preserving capabilities. It represents the separation of knowledge such that information available on the Demographic Public System1150A is not discoverable on the Demographic Private System1160B. Only the information that is explicitly shared from one side of the boundary to the other is available. For example, the relationship between the DeviceID and the DeviceToken is not discoverable on the Demographic Private System1160B inFIG. 11B.
• Continuing withFIG. 11B, the Device Demographic Loader1107B generates log records from the input DeviceToken, request time, and Demographic Keys. The output from this process is the Device Demographic Profile Logs1108B.
• The Device Demographic Processor1110B reads records from the Device Demographic Profile Logs1108B as well as specific values from the Location Demographic Attributes1109B based on the Demographic Keys previously logged. The information in the Location Demographics Attributes1109B represent the values for the demographic attributes, for example, the ratio of people in the given area which are females based on previously known information. All Device Demographic Profile Logs records are gathered for each DeviceToken. Using this set of information, the Device Demographic Processor1110B will perform various clustering algorithms to compute the most likely set of demographics that represent the Mobile Device for the specified criteria. For example if the ‘Home’ demographic for the device is requested, the cluster algorithm may choose all demographic log records that occurred between 10 PM and 5 AM to represent the most likely times that a device would be located within the home demographics of the device owner.
• Clustering is a common data mining method by which groups of ‘like’ items are assigned to a set by computing a difference, or ‘distance’ between items. Cluster analysis can be accomplished with a number of different algorithms based on the specific need. In particular, a cluster of similar demographic records can be found by measuring the difference between two demographic records. Records with similar values can be clustered together. An example would be to measure the difference between all attribute's buckets. If the average value between buckets is less than 0.1, these records would be considered as part of the cluster.
• Upon successful clustering by the Device Demographic Processor1110B, a set of records that represent the computed demographics for the device will be written to theDevice Demographic Profiles111 lB database and a relationship between the DeviceToken and the DemographicProfileToken will be stored in the DemographicTokens1126B database.
• Continuing inFIG. 11B, a copy of the Demographic Tokens1126B database is transferred to the Demographic Public System1150 inFIG. 11A (at1126A). TheDevice Demographic Profiles111 lB are optionally copied to the Demographic Public System1150A (at1111A) for use in computing a public version the Device Demographic Summary.
• Returning toFIG. 11A, a request by the Mobile Device1101 will also instigate a process within the Location Demographic Counter1114A. This process counts the presence of unique devices within a given geographic boundary during a specific time window. For example, it may count the number of devices that appeared in a city block on Monday between 3 PM and 4 PM. Utilizing, for example, a Probabilistic Counter1115A such as a Bloom Filter, the unique number of devices can be counted without identifying the specific Devices that were counted, thus preserving the privacy of the individual devices that are found to be present at various locations. These counts are then passed to the Device Count Loader1116B located within the administrative domain of the Demographic Private System1160B shown inFIG. 11B.
• The Device Count Loader1116B generates records that are stored in the Device Count Logs1117B database inFIG. 11B.
• Returning toFIG. 11A, a request by the Mobile Device1101 will also instigate a process within theLocation Demographic Logger1118A. This process will convert the DeviceID to a DeviceToken. It will then use this information to find a matching record, if available, within the DemographicTokens database. If present, the record will contain the DemographicToken for the given DeviceToken. The specific location is converted to a geometric shape representing some area larger than the identified point. This process reduces the specificity of the location, providing additional privacy preserving properties. Additionally, the specific time of the request is reduced in a manner similar to that described within the Device Demographic Logger1105A process to provide additional privacy preserving qualities. The DemographicToken, the location geometry, and the modified time value are then sent to the Location Demographic Loader1119B within the Demographic Private System1160B shown inFIG. 11B.
• Continuing inFIG. 11B, the Location Demographic Loader1119B generates records which are stored in the Location Demographic Logs1120B database.
• The Location Demographic Processor1121B periodically computes the Location Demographic Profiles for each location geometry for which the Location Demographic Logs1120B database contains information. For each location geometry, the process retrieves records for a specific time frame (e.g. each hour). These records are combined with theDevice Demographic Profiles111 lB database to compute the aggregate Location Demographic Profile for that location during that time duration.
• For each Location Demographic Logs1120B record, the associated Device Demographic Profiles1111B record is retrieved based on the Demographic Token. To combine the demographic information, a number of techniques can be applied. For example, the median of the demographics for each attribute can be chosen. Using median removes outliers and considers the best representative of temporal/spatial demographics. The median can be modified to select a bin with the most number of like records, if they exist. In this case, the most frequently occurring values could be chosen as the result for each attribute and bucket.
• An alternative method would be to use the average of demographics, which is calculated by averaging each bucket of each attribute. Other statistical means to compute the combined demographics could be applied (e.g. weighted average based on number of like samples).
• The Device Count Logs1117B are consulted to assist in computing the confidence in the profile based on the number of devices observed within the given location geometry during the time in question. The confidence can be computed based on the total number of samples observed for that location within the time window. This number can be compared to historical sample counts to determine if the sample count is statistically significant. Alternatively, the expected population of the area can be used to compare the sample count to determine the statistical significance of the samples. In one embodiment, the statistical significance would be used as the confidence factor. Other factors, such as the accuracy of the location, or externally available information about the number of mobile device users in the area could be used to compute the confidence factors.
• This process is repeated for each demographic attribute and for each desired time span. This results in record for each computation that includes the location geometry, the time span, the confidence factor(s) and values for each computed bucket within the given attribute (e.g. the computed ratio of males and females). These records are then written to theLocation Demographic1112B database.
• TheLocation Demographic database1112B is copied to the Demographic Public System1150A depicted inFIG. 11A.
• Returning toFIG. 11A, if theMobile Device1101A has requested a Location Demographic Profile, the Location Demographic Summary1123A process is executed for the location that is requested. This process combines information from the Demographic Bucket Reference1125A database, the Demographic Attribute Reference1124A database, and the Location Demographics1112A database to generate a record indicating the computed demographics for a given geographic geometry for the desired time window. The reference tables allow the generic information contained in the Location Demographics1112A database to be converted back into semantically meaningful demographic attribute and bucket names (e.g. convert from “Table A” to “Education”, and “Bucketl” to “Bachelors”). This record is then returned to theMobile Device1101A.
• Optionally, theMobile Device1101A may request its own Device Demographic Summary. If requested, the Device Demographic Summary1113A process will be executed. This process will convert the DeviceID to a DeviceToken. Using this information, it will retrieve the Device Demographic Profiles1111A for this device. It will use this information as well as information obtained from the Demographic Tokens1126A database, the Demographic Bucket Reference1125A database and the Demographic Attribute Reference1124A database to generate a record that summarizes the computed demographics for the given device. This record is then returned to theMobile Device1101A. Note that providing this optional service reduces the privacy preserving nature of the system. In order to produce the Device Demographic Summary, the Demographic Public System1150A must have access to the Device Demographic Profiles1111A database, thus providing a potential for the exposure of this private information.
• FIG. 12 shows an embodiment of theprocess1200 to convert demographic data into privacy preserving subsets. This process occurs prior to the system being deployed to the Operational Data1230, and can be processed once for each demographic attribute that should be included in the demographic computations.FIG. 12 shows a specific demographic attribute, the Education Demographics1215, being converted from the Original Data1210 into its constituent parts in the Location Demographic Attributes1240 database and the Location Demographic Buckets1250 databases.
• Continuing withFIG. 12, the Demographic Encoder1220 reads the Education Demographics1215 database. For each record, it generates a reference ID, called the Recno. This Recno will serve as the link between the G1 Demographics1241 database and the A1 Demographics database1252. The Demographic Encoder1220 then stores the Location from the Education Demographics1215 record along with the generated Recno in the G1 Demographics database. It also stores the generated Recno along with the values from the columns in the Education Demographics1215 database into their respective columns in the A1 Demographics1252 database. For example, the values in the High School column from the Education Demographics1215 database will be stored in the Bucket 1 column of the A1 Demographics1252 database for their respective rows. The extent of the Locations in the Education Demographics1215 database can be any size and shape that provides a meaningful designation of likely common demographics shared by those within the extent. For example, the Locations can be a neighborhood, a city block, a venue, a small town, or anything on the order thereof.
• In addition to splitting the data values between two distinct tables, the relationship between these tables and new column names must be maintained in order to reverse the mapping when processed data is to be returned to the Mobile Device. To encode this mapping, two additional databases are required. The Demographic Attribute Reference1243 database encodes the mapping from the new table name (e.g. “A1”) to the original data table name (e.g. “Education”). The second database, Demographic Bucket Reference1242, provides the mapping for each bucket or column within an attribute. For example, this allows the system to convert from “Bucket 1” to the semantically meaningful value of “High School”.
• Continuing withFIG. 12, note that the division of the data is also maintained across theAdministrative Boundary1127. The G1 Demographics1241 database is contained within the Demographic Public System1150 while the A1 Demographics1252 is contained on the Demographic Private System1160. This division is an example that provides additional privacy preserving properties both by obfuscating the information (e.g. converting recognizable information such as “High School” to generic labels such as “Bucket 1”) as well as separation of knowledge (e.g. the relationship of Recno to Area is only known by the Demographic Public System1150, while the relationship of Recno to attribute buckets is only known by the Demographic Private System1160.
• FIG. 13A andFIG. 13B show an instance ofFIG. 11A andFIG. 11B containing additional detail with respect to the data elements used for privacy preserving purposes. In particular,FIG. 13A andFIG. 13B depict an embodiment of the system in which the abstract database icons and names fromFIG. 11A andFIG. 11B have been replaced with specific instances of tables and data to provide more concrete examples of the privacy preserving features. In particular, it depicts a single demographic attribute, “Education” and the attribute buckets of “High School”, “Bachelors”, and “PhD”.
• InFIG. 13A, the Mobile Device1301A requests either a Location or Device Demographic Profile or a combination of both from a Location andDemographic Service1300A. The Mobile Device1301A passes either the location of the device or enough information for the service to compute its location in conjunction with a DeviceID.
• TheDemographic Service1300A receives the request and routes it to theDemographic Public System1350A which determines if location information is present via1302A. If location information is present, the request is routed to the Demographic Service1304A, otherwise the request is routed to the Location Service1303A. If the location is not present, the Location Service1303A will attempt to compute the location of the device. If it cannot determine the location of the device, an empty (or equivalent) response is returned to the Mobile Device1301A. Otherwise, the request is forwarded to the Demographic Service1304A along with the newly computed location. Henceforth, all activities continue as if the request had come from the Mobile Device1301A with location included.
• Continuing withFIG. 13A, the information is passed to three independent processes: the Device Demographic Logger1305A, the Location Demographic Counter1314A, and the Location Demographic Logger1318A.
• The Device Demographic Logger1305A provides a number of privacy preserving actions. It converts the DeviceID into a DeviceToken. It also converts the time of request to the local time based on the location of the device. This local time, in certain embodiments, may then be further reduced in specificity, for example by reducing it to indicate only hour of the week information. The Device Demographic Logger1305A then retrieves the demographic keys for the given location from the G1 Demographics1306A database. For example, assume the request was received by a location within “Area1”, the resulting key (e.g. Recno) would be “8230”. This Recno, the converted request time, and the DeviceToken are then passed over anAdministrative Boundary1327A to the Device Demographic Loader1307B shown inFIG. 13B.
• TheAdministrative Boundary1327A provides additional privacy preserving capabilities. It represents the separation of knowledge such that information available on theDemographic Public System1350A is not discoverable on the Demographic Private System1360B inFIG. 13B. Only the information that is explicitly shared from one side of the boundary to the other is available. For example, the relationship between the DeviceID and the DeviceToken is not discoverable on the Demographic Private System1360B.
• Referring toFIG. 13B, the Device Demographic Loader1307B generates log records from the input DeviceToken, request time, and Demographic Keys. The output from this process is the A1 Device Logs1308B. Each record contains the DeviceToken, the HourOfWeek, and the Recno as sent by the Device Demographic Logger1305A fromFIG. 13A.
• Continuing withFIG. 13B, the Device Demographic Processor1310B reads records from the A1 Device Logs1308B as well as specific values from the A1 Demographics1309B based on the Demographic Keys previously logged. The information in the A1 Demographics1309B represent the values for the buckets within the “A1” demographic attribute. All Device Demographic Profile Logs records are gathered for each DeviceToken. Using this set of information, the Device Demographic Processor1310B will perform various clustering algorithms to compute the most likely set of demographics that represent the Mobile Device for the specified criteria. For example if the ‘Home’ demographic for the device is requested, the cluster algorithm may choose all demographic log records that occurred between 10 PM and 5 AM to represent the most likely times that a device would be located within the home demographics of the device owner.
• Upon successful clustering by the Device Demographic Processor1310B, a set of records that represent the computed demographics for the device will be written to the Device Demographic Profiles1311B database and a relationship between the DeviceToken and the DemographicProfileToken will be stored in the Demographic Token1326B database. Examples of this output are depicted in the respective databases inFIG. 13B. In particular, it is noted that DeviceToken “ax8778as02” is connected to the Recno “8230” via this relationship as indicated inFIG. 13B.
• Continuing inFIG. 13B, a copy of the Demographic Token1326B database is transferred to theDemographic Public System1350A depicted onFIG. 13A (as1326A). The Device Demographic Profiles1311B database is optionally copied to theDemographic Public System1350A, shown inFIG. 13A, for use in computing a public version the Device Demographic Summary (as1311A).
• Returning toFIG. 13A, a request by the Mobile Device1301A will also instigate a process within the Location Demographic Counter1314A. This process counts the presence of unique devices within a given geographic boundary during a specific time window. For example, it may count the number of devices that appeared in a city block on Monday between 3 PM and 4 PM. Utilizing, for example, a Probabilistic Counter1315A such as a Bloom Filter, the unique number of devices can be counted without identifying the specific Devices that were counted, thus preserving the privacy of the individual devices that are found to be present at various locations. These counts are then passed to the Device Count Loader1316B located within the administrative domain of the Demographic Private System1360B depicted inFIG. 13B.
• Continuing withFIG. 13B, the Device Count Loader1316B generates records that are stored in the Device Count Logs1317B database. For example, referring toFIG. 13B that for Geometry “ABAF007” andhour 57 of the week, there were 8 distinct devices counted.
• Returning toFIG. 13A, a request by the Mobile Device1301A will also instigate a process within the Location Demographic Logger1318A. This process will convert the DeviceID to a DeviceToken. It will then use this information to find a matching record, if available, within the DemographicTokens database1326A. If present, the record will contain the DemographicToken for the given DeviceToken. The specific location is converted to a geometric shape representing some area larger than the identified point. This process reduces the specificity of the location, providing additional privacy preserving properties. Additionally, the specific time of the request is reduced in a manner similar to that described within the Device Demographic Logger1305A process to provide additional privacy preserving qualities. The Demographic Token, the location geometry, and the modified time value are then sent to the Location Demographic Loader1319B within the Demographic Private System1360B shown inFIG. 13B.
• Continuing inFIG. 13B, the Location Demographic Loader1319B generates records which are stored in the Location Demographic Logs1320B database.
• The Location Demographic Processor1321B periodically computes the Location Demographic Profiles for each location geometry for which the Location Demographic Logs1320B database contains information. For each location geometry, the process retrieves records for a specific time frame (e.g. each hour). These records are combined with the Device Demographic Profiles1311B database to compute the aggregate Location Demographic Profile for that location during that time duration. The Device Count Logs1317B are consulted to assist in computing the confidence in the profile based on the number of devices observed within the given location geometry during the time in question. This process is repeated for each demographic attribute and for each desired time span. This results in a record for each computation that includes the location geometry, the time span, the confidence factor(s) and values for each computed bucket within the given attribute. These records are then written to the A1 Location Demographics1312B database. Referring toFIG. 13B, there is a single entry in A1 Location Demographics1312B database representing the computed values for demographic attribute “A1” at location geometry “ABAF007” for the 57th hour of the week.
• The A1 Location Demographics database1312B is copied to the Demographic Public System1350B (as1312A) shown inFIG. 13A.
• Returning toFIG. 13A, if the Mobile Device1301A has requested a Location Demographic Profile, the Location Demographic Summary1323A process is executed for the location that was requested. This process combines information from theDemographic Bucket Reference1325A database, the Demographic Attribute Reference1324A database, and the Location Demographics1312A database to generate a record indicating the computed demographics for a given geographic geometry for the desired time window. The reference tables allow the generic information contained in the A1 Location Demographics1312A database to be converted back into semantically meaningful demographic attribute and bucket names. This record is then returned to the Mobile Device1301A.
• FIG. 14 depicts anexample output1400 based for an Location Demographic Profile produced by the system shown inFIG. 13A andFIG. 13B. This represents a partialdemographic profile1401 based on the limited example data contained in the figure.
• Optionally, the Mobile Device1301A may request its own Device Demographic Summary. If requested, the Device Demographic Summary1313A process will be executed. This process will convert the DeviceID to a DeviceToken. Using this information, it will retrieve the Device Demographic Profiles1312A for this device. It will use this information as well as information obtained from the Demographic Token1326A database, theDemographic Bucket Reference1325A database and the Demographic Attribute Reference1324A database to generate a record that summarizes the computed demographics for the given device. This record is then returned to the Mobile Device1301A. Note that providing this optional service reduces the privacy preserving nature of the system. In order to produce the Device Demographic Summary, theDemographic Public System1350A must have access to the Device Demographic Profiles1311A database, thus providing a potential for the exposure of this private information.
• In the foregoing description, certain steps or processes were described as being performed on particular servers or as part of a particular engine. These descriptions are merely illustrative, as the specific steps can be performed on various hardware devices, including, but not limited to, server systems and/or mobile devices. Similarly, the division of where the particular steps are performed in the above description illustrates certain embodiments, if being understood that no division or a different division is within the scope of the invention.
• The techniques and systems disclosed herein may be implemented as a computer program product for use with a computer system or computerized electronic device. Such implementations may include a series of computer instructions, or logic, fixed either on a tangible medium, such as a computer readable medium (e.g., a diskette, CD-ROM, ROM, flash memory or other memory or fixed disk) or transmittable to a computer system or a device, via a modem or other interface device, such as a communications adapter connected to a network over a medium.
• The medium may be either a tangible medium (e.g., optical or analog communications lines) or a medium implemented with wireless techniques (e.g., Wi-Fi, cellular, microwave, infrared or other transmission techniques). The series of computer instructions embodies at least part of the functionality described herein with respect to the system. Those skilled in the art should appreciate that such computer instructions can be written in a number of programming languages for use with many computer architectures or operating systems.
• Furthermore, such instructions may be stored in any tangible memory device, such as semiconductor, magnetic, optical or other memory devices, and may be transmitted using any communications technology, such as optical, infrared, microwave, or other transmission technologies.
• It is expected that such a computer program product may be distributed as a removable medium with accompanying printed or electronic documentation (e.g., shrink wrapped software), preloaded with a computer system (e.g., on system ROM or fixed disk), or distributed from a server or electronic bulletin board over the network (e.g., the Internet or World Wide Web). Of course, some embodiments of the invention may be implemented as a combination of both software (e.g., a computer program product) and hardware. Still other embodiments of the invention are implemented as entirely hardware, or entirely software (e.g., a computer program product).
• Moreover, the techniques and systems disclosed herein can be used with a variety of mobile devices. For example, mobile telephones, smart phones, personal digital assistants, satellite positioning units (e.g., GPS devices), and/or mobile computing devices capable of receiving the signals discussed herein can be used in implementations of the invention.

Claims (23)

1. A method of estimating demographic information associated with a user of a mobile device while preserving the privacy of the user based at least in part on a location estimate of the mobile device of the user, the method comprising:

(a) receiving an estimated geographical location of the mobile device of the user;

(b) receiving a time at which the mobile device was at the estimated geographical location;

(c) providing a set of substitute identifiers for a corresponding set of at least one geographical area;

(d) assigning one of the set of substitute identifiers for the geographical area corresponding to the geographical location of the mobile device;

(e) assigning a substitute identifier for the time at which the mobile device was at the estimated geographical location;

(f) providing an association between the substitute identifiers for geographical areas and demographic information corresponding to the substituted geographical area; and

(g) estimating demographic information associated with the user of the mobile device based on the assigned substitute identifiers and based on the demographic information associated with the provided set of substitute identifiers.

2. The method ofclaim 1, further comprising (h) estimating the received geographical location associated with the mobile device of the user.

3. The method ofclaim 1, further comprising (h) recording the estimated demographic information associated with the user of the mobile device in a device demographic information log.

4. The method ofclaim 3, further comprising (i) estimating demographic information associated with at least one geographical area of the set of geographical areas based on the demographic information recorded in the device demographic information log, wherein the device demographic information log contains a plurality of records of estimated demographic information associated with a plurality of mobile devices.

5. The method ofclaim 1, further comprising (h) sending the estimated demographic information to the mobile device.

6. The method ofclaim 1, wherein the substitute identifier for the geographical area corresponding to the geographical location of the mobile device that is assigned identifies a particular set of demographic information.

7. The method ofclaim 1, wherein the substitute identifier for the geographical area corresponding to the geographical location of the mobile device that is assigned is reduced in specificity relative to the estimated geographical location of the mobile device.

8. The method ofclaim 1, wherein the substitute identifier for the time at which the mobile device was at the estimated geographic location is a measure of time that is reduced in specificity relative to the time that was received.

9. The method ofclaim 8, wherein the substitute identifier for the time at which the mobile device was at the estimated geographic location is at least one of (A) a representation of time lacking date information, (B) a time range, and (C) an hour of a week designation.

10. The method ofclaim 1, wherein the estimating demographic information associated with the user of the mobile device is performed on a separate computer system from a computer system performing any one or more of steps (a) through (e).

11. The method ofclaim 10, wherein the separate computer system, relative to the computer system performing any one or more of steps (a) through (e), is as least one of (A) maintained in a separate network, (B) maintained in a separate building, and (C) maintained by a separate operational entity.

12. A method of estimating demographic information associated with a geographical area and a time period based on demographic information associated with users of mobile devices within the geographical area, the method comprising:

(a) providing a set of geographical areas;

(b) providing a set of time periods;

(c) receiving an estimated geographical location of a mobile device of the user;

(d) receiving a mobile device identifier that is associated with the mobile device;

(e) receiving a time at which the mobile device was at the estimated geographical location;

(f) determining the geographical area of the set in which the estimated geographical location occurs;

(g) determining the time period of the set in which the time at which the mobile device was at the estimated geographical location occurs;

(h) retrieving information representative of demographic information associated with the user of the mobile device based on the mobile device identifier;

(i) estimating demographic information associated with the determined geographic area during the determined time period based on the retrieved information representative of demographic information associated with the user of the mobile device.

13. The method ofclaim 12, further comprising (j) estimating the received geographical location of the mobile device of the user.

14. The method ofclaim 12, further comprising (j) providing a set of initial demographic information associated with the determined geographical area, wherein the estimating demographic information associated with the determined geographic area is further based on the initial demographic information.

15. The method ofclaim 14, further comprising, (k) subsequent to the estimating demographic information associated with the determined geographical area, adjusting the initial demographic information based on the estimated demographic information.

16. The method ofclaim 14, wherein the initial demographic information is based on at least one of governmental census information, public record information, shipping and purchase records, magazine and newspaper subscriptions, voluntary surveys, and records of social media activity.

17. The method ofclaim 12, further comprising (j) sending the estimated demographic information to the mobile device.

18. The method ofclaim 12, wherein the geographical areas of the set are reduced in specificity relative to the estimated geographical location of the mobile device.

19. The method ofclaim 12, wherein the time periods of the set are reduced in specificity relative to the time that was received.

20. The method ofclaim 19, wherein the time periods of the set are at least one of (A) a representation of time lacking date information, (B) a time range, and (C) an hour of a week designation.

21. The method ofclaim 12, wherein the estimating demographic information associated with the determined geographic area during the determined time period is performed on a separate computer system from a computer system performing any one or more of steps (c) through (g).

22. The method ofclaim 21, wherein the separate computer system, relative to the computer system performing any one or more of steps (c) through (g), is as least one of (A) maintained in a separate network, (B) maintained in a separate building, and (C) maintained by a separate operational entity.

23. The method ofclaim 12, further comprising:

(j) performing steps (c) through (h) a plurality of times for different mobile devices of different users; and

(k) recording sets of information for (A) the determined geographical areas, (B) determined time periods, and (C) retrieved information representative of demographics information associated with the users of the mobile devices in a device demographic information log;

wherein the estimating demographic information associated with the determined geographical area during the determined time period is further based on a plurality of the sets of information in the device demographic information log.

Images