| City: |
San Francisco, California, USA |
| Organization: |
California State University |
| Project Start Date: |
Unknown |
| Project End Date: |
January 2014 |
| Reference: |
Sullivan, B., & Mitra, S. (2014). Community Issues in American Metropolitan Cities. Journal of Cases on Information Technology, 16(1), 23-39. doi:10.4018/jcit.2014010103 City and County of San Francisco, Office of the Controller. (2013). City Survey Dataset, 1996 to 2013 Maslin, M. Metz & Associates. San Francisco city report 2013 executive summary. Sfpl.org. |
| Problem: |
The City of San Francisco has slow population growth rate. The California State University uses data mining processes to identify major factors that drive people moving out of the city. |
| Technical Solution: |
Heatmap: identify multicollinearity (correlation between independent variables) occurred in the study, combine or eliminate independent variable.
Supervised learning: 60% of the survey data is the training data and 40% of the survey data is used as testing data.
Logistic Regression: Data mining tool used to analyze binary data, can analyze one or more independent variables with respect to dependent variable to determine outcome.
K-Nearest Neighbours: classify a resident based on the nearest K number in dataset to compare with logistic regression model outcome.
|
| Datasets Used: |
- San Francisco's Biennial City Survey, City of San Francisco, 2011
|
| Outcome: |
Pre Solution Performance
Previous studies of survey data used average and frequency distributions of responses;
Post Solution Performance
Discovered four areas city should set priories when ranking city projects, these including improve quality of School, extensive public transit, improve city parks and high expanse of living.
|
| Issues that arose: |
survey was only conducted by phone, people who did not have telephone cannot be reached.
97.66% of people who will move away was incorrectly modeled, this is due to in the training data people who will move away is also low; therefore, it is hard for model to learn the patterns.
|
| Status: |
Terminated |
| Entered by: |
November 12, 2017, May Wang mayx.wang@mail.utoronto.ca.
|