hamishgibbs / colocation_dashboard

Interactive dashboard of Facebook colocation data

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

colocation_dashboard

Interactive dashboard of Facebook colocation data.

JavaScript files hosted at: hamishgibbs/colocation_dashboard_js.

Data updates: Colocation data is made available by Facebook weekly. The LSHTM colocation dashboard will be updated as soon as possible following a weekly release of colocation data.

Please note: These datasets have been aggregated from their original format in order to be released publicly.

To report a problem with this dashboard or the underlying data, please create an issue.

Data Reference

name_replacement.csv
Facebook data is referenced to administrative boundary datasets produced by Pitney Bowes. The datasets provided in this repository are referenced to publicly available GADM boundary datasets. In order to symbolise data using GADM boundaries, certain administrative names in the input data have been altered. This file records the mapping of administrative names from Pitney Bowes polygons to GADM polygons.

Fields:
fb_name: The administrative name provided by Facebook.
replacement: The GADM name for the same administrative area.

mean_ts.csv
The mean colocation probability between a given area and all other areas, each week.

Fields:
connection_type: Contains a single value - Between. Specifying that the record regards data between areas, not within the same area.
polygon1_name: The administrative name of the "target" polygon.
ds: The date of the record.
mean_colocation: The colocation probability (see dashboard for more details).
NAME_1: Country name.
GID_2: GADM formatted area code.
TYPE_2: Administrative area type.
type: Type of record, either perc_change or abs_value. Whether the record displays unaltered mean colocation probabilites (abs_value), or percent change from the first timeseries point (perc_change).

top_n_between.csv
The mean probability of colocation across the entire timeseries between selected pairs of locations. The 10 locations with the highest colocation probabilities are selected for each area.

Fields:
polygon1_name: The administrative name of the "target" polygon.
polygon2_name: The administrative name of the "connected" polygon.
mean_colocation: The mean probability of colocation across the entire timeseries between polygon1 and polygon2.

Please see the dashboard for further information about the colocation metric.

About

Interactive dashboard of Facebook colocation data


Languages

Language:R 54.5%Language:HTML 30.5%Language:Makefile 11.8%Language:JavaScript 3.1%