Social Philly

Using Twitter to Measure Activities and Sentiments


Subject of Study

This project aims to use Twitter to evaluate the quality of life of low-income housing tax credit (LIHTC) program receivers. The density level of tweets is evaluated to measure the activity level of different neighborhoods. Besides activity level, we also want to extract the emojis in each tweet to have a general understanding of people's emotional tendency across the city. The measurement of sentiment score for each emoji is based on an emoji library constructed in the paper "Sentiment of Emojis."[1]. From the analysis, we can identify several active neighborhoods in the city of Philadelphia, a sentiment heat map for the city, and also an evaluation of the LIHTC developments across the city.


Data and Method

20,320 geotagged Tweets scraped across Philadelphia over two weeks were used for analysis. The Tweets were categorized based on the land use type of their location, and the density of Tweets were calculated using number of Tweets divided by built floor area or population.



Team

Haijing Liu, MCP '18

Xinhui Li, MCP '18

Course project of MIT 11.s943 Spring 2017.

Most of Tweets Happened on Mixed-use Parcels



In this map, the left side bars represent the average score for each landuse types in every planning district. You can click on the bars to check the activity level and sentiment level of each planning district. On the right side, you can hover on each planning district to check their detailed information.



More Tweets Happened on LIHTC Developments than Other Residential Parcels



In this map, the left side bars represent the average score for normal residential developments and LIHTC developments in every planning district. You can click on the bars to check the activity level and sentiment level of each planning district. On the right side, you can hover on each planning district to check their detailed information.



    • Number of Tweets within 300ft