Rajesh Jagannath April 30, 2017
The Cincinnati Streetcar is a modern streetcar system designed to link major employment centers in downtown and uptown, connecting through Cincinnati's historic Over-the-Rhine neighborhood. It was introduced into service in August 2016. It operates 18 hours a day, 365 days a year.
Impacts on the economy were expected and anticipated upon its introduction - some positive and some negative.
This project analyzes Total Market Value of the parcels in a buffer zone around the Streetcar route. Three spatial buffer-zones were established by Cincinnati Area Geographic Systems(CAGIS). A random sample of parcels not in the Buffer-zone were compared. Polynomial Regression was used for prediction of MKT_TOTAL_VAL(Total Market Value). Data was used with permission from the staff at CAGIS. Using ClustofVars library, clusters of parcels_ids with similar trends are grouped for analysis into k=12 clusters. Training sets and test sets were created. Error (Mean Absolute Percentage Error) was minimized on Test set to select the best model for prediction. Models were selected from one of linear, quadratic, cubic, 4th order or 5th order polynomial regression. Prediction and Visualization were performed. Insights were drawn and report concluded with ideas for future work.
I have selected this study as my capstone project submission for the Springboard - Foundations to Data Science Course.
Please refer to the Final Project Report for the full details of the problem and dataset.
A power point presentation of the project exists Project Slides
__A output file of main.Rmd is located here __ main.html
| File | Description |
|---|---|
| taxinfoTimeSeries/main.R | Main R Script for ETL, Clustering, Model selection, Prediction and Visualization |
| taxinfoTimeSeries/Buffer_parcels_ETL_MKT_VAL.R | ETL files for the 3 buffer zone CORE,CENTER and EDGE |
| taxinfoTimeSeries/Assessors_yearly_data_ETL_MKT_VAL.R | ETL for loading 8 years of data |
| taxinfoTimeSeries/functions_MKT_VAL.R | Contains functions for Polynomial regression |
| Data2_090216/Read_Position_delimited_File_from_tuple.R | reads in the the txt files in fixed width format |
| Reports/Final_Milestone/FinalReport_041617.pdf | Final Report |
| Reports/Final_Milestone/Final_ReportPresentation.pdf | Final Power point Presentation |
| Directory | Description |
|---|---|
| Data2_090216 | Contains datasets |
| Data2_090216/parcel_csv | contains the parcel ids of the parcels in Buffer Parcel (csv files) |
| Data2_090216/taxinfo | contains the 9 years of data including MKT_TOTAL_VAL (txt file - fixed width format) |
| taxinfoTimeSeries | Contains the main code |
| Reports/Final_Milestone | The final reports are archived here |
| Reports/Final_Milestone/Plots | contains all the supporting plots generated in the study |
- https://dev.socrata.com/foundry/data.cincinnati-oh.gov/emnx-rw6d
- http://www.cincinnati.com/story/news/2016/05/05/streetcar-nation-kc-opens-friday-cincy-next/83874740/
- https://cran.r-project.org/web/packages/ClustOfVar/ClustOfVar.pdf
- http://www.exegetic.biz/blog/2013/12/contour-and-density-layers-with-ggmap/
- http://www.shanelynn.ie/massive-geocoding-with-r-and-google-maps/
- http://stat405.had.co.nz/ggmap.pdf
- A Little Book of R for Time-series, Avril Coghlan
- An Introduction to Statistical Learning, Gareth James et al