We will evaluate the impact of financial disparities on test scores in K-12 schools in Washington D.C. before and after the COVID-19 pandemic. We want to find out whether students' academic performances been affected by the varying levels of financial support that each school can apply.
We hypothesize that a school's level of spending plays a factor in how much test scores varied; meaning that the more resources the school was able to provide, the less the math and reading scores dropped from 2019 to 2021.
Data Sources (Google Drive)
- For measuring students' academic performances we use NAEP data
- For school spending data we use Edunomics Lab data
- For demographic data of children and parents we use NCES data
- For geographic data of school district we use Census data
- Which DC schools had the biggest drop-offs from 2019 to 2022?
- Where were students best able to adapt to online learning?
- How the family income levels and the financial disparities among schools contribute to students' adaptability to online learning?
- How the pandemic impact differently on DC's public schools and charter schools?
000_ pullNCESdata: https://github.com/shiuanC/final_team7/blob/main/code/00_pullNCESdata.ipynb
- takes in an excel file containing the various school's financial data form 2018-2019 and 2019-2020
- uses nces_process function to clean the two datasets
- merges the two academic year dataframes into one nces dataframe
- calculates the percentage change in per pupil expenditure from 2018-2019 to 2019-2020
- returns clean csv dataframe called nces
001_pullSTARdata: https://github.com/shiuanC/final_team7/blob/main/code/01_pullSTARdata.ipynb
- takes in an excel file containing data on each school's STAR reports from 2018-2019 and 2019-2020
- cleans column names
- merges the two academic year dataframes into one star dataframe
- calculates the percentage change in STAR score from 2018-2019 to 2019-2020
- returns clean csv dataframe called star
002_pullDiversityData: https://github.com/shiuanC/final_team7/blob/main/code/03_PullDiversityData.ipynb
- takes in an excel file containing data on student background's, msot importantly the percentage of student's deemed "at-risk", from 2018-2019 and 2019-2020
- gives each school a label on a scale of 1-5 according to risk percentage
- schools labeled 1 having a small number of "at-risk" students and schools labeled 5 having a lot of "at-risk" students
- returns clean csv dataframe called diversity
003_MergeDataframes: https://github.com/shiuanC/final_team7/blob/main/code/04_MergeAllDataframes.ipynb
- takes in nces, star, and diversity csv files
- merges the dataframes together on school code
- only keeps school's that has both STAR and NCES data
- cleans columns
- returns clean csv file called all_df
004_Diagnosis_missing_schools: http://localhost:8889/notebooks/code/004_Diagnosis_missing_schools.ipynb
- takes in each dataset
- evaluates missing values
- finds duplicate columns
- nalayzing which schools were dropped during the merging process
005_DescriptiveFigs: https://github.com/shiuanC/final_team7/blob/main/code/05_DescriptiveFigs.ipynb
- takes in all_df csv file
- returns plot of percentage of at risk students against total per-pupil ecxpenditures
- returns plot of change in STAR scores against change in per puil expenditures by percentage of "at-risk" students
- returns plot of the change in per puil expenditure and change in STAR score for each DC public school
- returns plot of average change in STAR score for each ward
- returns plot of average change in STAR score and average per-pupil expenditure for each ward
- returns plot of each school's STAR score, per-pupil expenditure, at-risk percentage, and enrollment size
- returns seperate geoplots for each school's STAR score and per-puil expenditure
006_CorrelationRegression: https://github.com/shiuanC/final_team7/blob/main/code/06_CorrelationRegression.ipynb
- takes in all_df csv file
- uses an ols regression model to see the effect of per-pupil expenditures STAR score
- returns summary table of regression findings