Better or Worse?

GitHub link

Problem

School redistricting in Linganore-Oakdale-Urbana (LOU) area has stirred outcry from the local communities in the social media. In order to understand the concerns about local communities, Frederick County Board of Education provided an online platform for community members to express their opinions and conducted a basic statistics of local community concerns. However, their analyses lack of details about what sentiment of the local communities were and why the communities show favor/disfavor to the school redistricting plans.

Insights

• The LOU commmunities show more positive attitudes towards the second-round proposed school redistricting plans. Particularly, both Options A and B of high school districts have much less negative feedbacks than previous versions. Local communities that have childrens enrolled in Linganore HS prefer Option B instead of Option A; those in Oakdale HS like Option A more than Option B. Urbana HS parents seems OK for both options.
• For middle school districts, communities that have their children enrolled in New Market MS, Oakdale MS, Urbana MS largely prefer Option A. Windosor Knolls MS community shows neurtual attitde towards both options, despite its sentiment score of the Option B is slightly higher. Some parents of Thomas Johnson MS show great objection to the Option A, but relatively neutral attitude to the Option B.
• Among elementary school districts, the Option A of Deer Crossing ES, Centerville ES, and Spring Ridge ES were largely preferred by local communities, while the Option B gain positive feedbacks among parents of Spring Ridge ES, Green Valley, and Kempton ES. It seems that Spring Ridge ES community particularly favored the Option B.
• Using interactive sentiment analyses, local education administration could effectively identify where local community needs were and tailor its resources to address them, which would save time and money.
• My web app could serve as a platform for general public to understand their neighbors’ sentiments of the school redistricting studies.

Data

The community feedbacks for the second-round proposals of new school districts were assembled in single pdf file that contains numerous tables stretching over 200-page long. The tables store feedbacks for elementary, middle, and high school districts of LOU area in Frederick County, MD. Each school district have tables for options A and B as well as some comments related but not specific to any option. I designated those unspecific comments to option AB.

Approach

In order scrape the data from the pdf file, I extract the comments and parse them into dataframe and stored them in a SQL database. To improve the efficiency of analyzing the data, I created redistrict module to scrape data, to perform exploratory data analyses and sentiment analyses, and to finally visualize the insights.

Tools and Techs

• Python
  • Camelot
  • Worldcloud
  • Streamlit
  • Folium
  • GeoPandas
  • NLTK
  • Plotly
  • Vincent
  • Geopy
• OpenStreetMap(OSM)
• NLP
• Feature Engineering
• SQL
• AWS EC2, S3

back