Just sharing stuff…
In the Philippine map below, we show a month’s worth of geotagged tweets collected via Twitter’s Public Streaming API spritzer, which allows authenticated users to access 1% of all public tweets worldwide. Two things immediately came to mind after I generated the image: (1) the correlation of the tweet distribution to the Gross Regional Domestic Product (GRDP), which is just the GDP measured at regional levels, and (2) the correlation of the distribution to Population Density.
Is it possible to just look at the number of tweets per region and infer the population density at that region? Does it make sense to obtain a rough estimate of the GRDP of a locality based on the number of geotagged tweets that fall onto the area?
To explore these ideas, I show two heat maps below: population density and GRDP. At first glance, that may actually seem to be the case—that Twitter data can be used as proxy for GRDP and population density (at least for the Philippines). However, the hypotheses need further investigation/validation since the heat maps show the trends for the whole of each region, but the PH Twitter map presents a much higher resolution visualization. Nevertheless, the results look encouraging. Hence, the thing I did next was to normalize the tweet counts per region—by doing so, we can finally compare apples with apples.
What else can the Twitter data represent? I have often told friends and colleagues that Twitter may not be a good survey source especially for a developing country like the Philippines where there is an existing issue on the penetration rate of social media platforms. But, if more citizens use Twitter or Facebook, does it mean that the city is technologically more “advanced” (in a relative sense)? Can it represent technological advancement? How about urbanization?
Do more urbanized cities have more dynamic Twitter activities?
Then I remembered NASA’s city light data from its Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS). The city light data have been used to map urbanization—where “brighter” regions (not necessarily more populated ones) have been identified to be more urbanized. The thought got me really excited!
Below, the Philippine Twitter map I generated (left) is placed side-by-side with the map of the Earth’s city light capture (right, zoomed in on the Philippines) .
Indeed, it looks like the Twitter data can represent urbanization, and it’s beautiful!
I then focused on the three Philippine metropolises that I thought are more urbanized than the rest of the Philippines: Metro Manila, Metro Cebu, and Metro Davao. The first two rows show the road networks of the metros—the only difference is in the rendering; on the other hand, the last row shows the mapped tweets. We can see from the figures that for Cebu and Davao, the tweets are concentrated within the mall areas—two things: (1) free wifi and (2) activity/event that’s worth tweeting about.
I find it interesting that there also seems to be a correlation between the intricacy of the road networks and the number of tweets. This is definitely worth investigating!
In the image below, I added the rail/rapid transit systems stations (purple) including some stations of the PNR.
Some idea on transport planning: Maybe the generated Twitter map can also be used as a guide, at least for Metro Manila, in extending/expanding our rapid/rail transit systems, e.g. Where should we place the interchanges?
 Data courtesy Marc Imhoff of NASA GSFC and Christopher Elvidge of NOAA NGDC. Image by Craig Mayhew and Robert Simmon, NASA GSFC. Date Data: October 1, 1994 – March 31, 1995
Legara E. Urbanism in the Philippines. A Byte of my 22-lb Brain. 2015. Available at: http://erikafille.ph/2015/09/10/urbanism-in-the-philippines/. Accessed Month DD, YYYY.