Computer system with features for determining reliable location data using messages with unreliable location data
Reliable location data is found using messages with unreliable location data. A logistic-geohash is found for a plurality of messages. Records of are accessed data messages, each data message including a geographic locator and a sender-identifier. For each sender-identifier, one or more geogroups are determined. For each geogroup: determining, a representative point of the geogroup is found to represent the geogroup. Using the representative point of the geogroup, a geogroup-geohash is found. Pair of geogroup: logistic address are found based on a matching of at least some of the first plurality of digits and some of the second plurality of digits. For each pair of geogroup:logistic address, a distance is determined between the representative point of the geogroup and the representative point of the logistic address. Geogroup:logistic address pairs are selected. The selected geogroup:logistic address pairs are stored in memory as representing a logistic location of the sender-identifier.
1. A system for determining reliable location data using messages with unreliable location data, the system comprising:
one or more processors;
memory storing instructions that, when executed by the processors, cause the processors to perform operations comprising:
determining, for each of a plurality of logistic addresses using a representative point of the logistic address, a logistic-geohash that comprises a first plurality of digits, the digits being arranged in sequence identifying progressively more granular geographic areas;
accessing records of data messages, each data message including a geographic locator and a sender-identifier;
for each sender-identifier, determining one or more geogroups, each geogroup identifying data messages originating near each other geographically;
for each geogroup:
determining a representative point of the geogroup to represent the geogroup;
determining, using the representative point of the geogroup, a geogroup-geohash that comprises a second plurality of digits, the digits being arranged in sequence identifying progressively more granular geographic areas;
identifying pairs of geogroup:logistic address based on a matching of at least some of the first plurality of digits and some of the second plurality of digits;
for each pair of geogroup:logistic address, determining a distance between the representative point of the geogroup and the representative point of the logistic address;
selecting one or more of the geogroup:logistic address pairs; and
storing in the computer memory, for later access, the selected geogroup:logistic address pairs as representing a logistic location of the sender-identifier of the messages of the geogroup.
2. The system of claim 1 , wherein the sender-identifier is one of the group comprising internet protocol (IP) address, digital household identifier, and user identifier.
3. The system of claim 1 , wherein determining the geogroups comprises identifying connected components in a graph created using geographic locators as vertices and distances between geographic locators as edges.
4. The system of claim 3 , wherein the graph is assembled using only geographic locators within a threshold distance of a landmark.
5. The system of claim 1 , wherein determining the geogroups comprises identifying clusters using a clustering analysis.
6. The system of claim 1 , wherein the representative points of geogroups are centroids of a shape containing the geographic locators of the corresponding identified data messages.
7. The system of claim 6 , wherein the centroids are found using a marginal median analysis.
8. The system of claim 1 , wherein the logistic address is one of the group comprising a mailing address, a building identifier, and a land parcel identifier.
9. The system of claim 1 , wherein identifying pairs of geogroup:logistic address based on a matching of at least some of the first plurality of digits and some of the second plurality of digits comprises:
generating a result table by joining a first table and a second table on N most significant digits in geohashes, the geohashes comprising more than N digits;
the first data table comprising entries that each record at least a logistic address and an associated logistic-geohash; and
the second data table comprising entries that each record at least a geogroup identifier and an associated geogroup-geohash.
10. The system of claim 1 , the operations further comprising initiating a computational process for a user associated with a particular sender-identifier using at least one of the selected geogroup:logistic address pairs as representing a logistic location of the sender-identifier of the messages of the geogroup.
11. A method of determining reliable location data using messages with unreliable location data, the method comprising:
determining, for each of a plurality of logistic addresses using a representative point of the logistic address, a logistic-geohash that comprises a first plurality of digits, the digits being arranged in sequence identifying progressively more granular geographic areas;
accessing records of data messages, each data message including a geographic locator and a sender-identifier;
for each sender-identifier, determining one or more geogroups, each geogroup identifying data messages originating near each other geographically;
for each geogroup:
determining a representative point of the geogroup to represent the geogroup;
determining, using the representative point of the geogroup, a geogroup-geohash that comprises a second plurality of digits, the digits being arranged in sequence identifying progressively more granular geographic areas;
identifying pairs of geogroup:logistic address based on a matching of at least some of the first plurality of digits and some of the second plurality of digits;
for each pair of geogroup:logistic address, determining a distance between the representative point of the geogroup and the representative point of the logistic address;
selecting one or more of the geogroup:logistic address pairs; and
storing in computer memory, for later access, the selected geogroup:logistic address pairs as representing a logistic location of the sender-identifier of the messages of the geogroup.
12. The method of claim 11 , wherein the sender-identifier is one of the group comprising internet protocol (IP) address, digital household identifier, and user identifier.
13. The method of claim 11 , wherein determining the geogroups comprises identifying connected components in a graph created using geographic locators as vertices and distances between geographic locators as edges.
14. The method of claim 13 , wherein the graph is assembled using only geographic locators within a threshold distance of a landmark.
15. The method of claim 11 , wherein determining the geogroups comprises identifying clusters using a clustering analysis.
16. The method of claim 11 , wherein the representative points of geogroups are centroids of a shape containing the geographic locators of the corresponding identified data messages.
17. The method of claim 16 , wherein the centroids are found using a marginal median analysis.
18. The method of claim 11 , wherein the logistic address is one of the group comprising a mailing address, a building identifier, and a land parcel identifier.
19. The method of claim 11 , wherein identifying pairs of geogroup:logistic address based on a matching of at least some of the first plurality of digits and some of the second plurality of digits comprises:
generating a result table by joining a first table and a second table on N most significant digits in geohashes, the geohashes comprising more than N digits;
the first data table comprising entries that each record at least a logistic address and an associated logistic-geohash; and
the second data table comprising entries that each record at least a geogroup identifier and an associated geogroup-geohash.
20. The method of claim 11 , the method further comprising initiating a computational process for a user associated with a particular sender-identifier using at least one of the selected geogroup:logistic address pairs as representing a logistic location of the sender-identifier of the messages of the geogroup.