top of page
![CertificateBackground.png](https://static.wixstatic.com/media/55f531_d6679d5b06e14c81ae07bacb53692d5f~mv2.png/v1/fill/w_714,h_536,al_c,q_90,usm_0.66_1.00_0.01,enc_auto/CertificateBackground.png)
Certificate of Completion
![ErdosHorizontal.png](https://static.wixstatic.com/media/55f531_5a3b8885620c4f25b2d3edeca3ae2158~mv2.png/v1/fill/w_351,h_40,al_c,q_85,usm_0.66_1.00_0.01,enc_auto/ErdosHorizontal.png)
THIS ACKNOWLEDGES THAT
HAS COMPLETED THE FALL 2023 DATA SCIENCE BOOT CAMP
Souparna Purohit
![clear.png](https://static.wixstatic.com/media/55f531_b9f3f13ce3aa4af78af2cc6d3563b81b~mv2.png/v1/fill/w_3,h_3,al_c,lg_1,q_85,enc_auto/clear.png)
Roman Holowinsky, PhD
DECEMBER 07, 2023
DIRECTOR
DATE
TEAM
ExecutiveHour
Rouzbeh Modarresi Yazdi, Nicolas Fortier, Souparna Purohit, Irem Altiner
![clear.png](https://static.wixstatic.com/media/55f531_b9f3f13ce3aa4af78af2cc6d3563b81b~mv2.png/v1/fill/w_3,h_3,al_c,lg_1,q_85,enc_auto/clear.png)
Day-ahead price forecasting for New York City energy markets using standard time-series features (prices at previous time steps) as well as some exogenous data (natural gas prices in the area, weather-related features, etc). We trained and compared ARIMA, Dense NN, Convolutional NN, LSTMs and XGBoost regressors against a reasonable baseline model. These models could eventually be expanded and used by industrial entities to determine whether current Day-Ahead prices are advantageous compared to expected Real-Time prices 24 hours in the future.
![](https://static.wixstatic.com/media/a994932411404ef3bb797ba005125f5d.png/v1/fill/w_45,h_45,al_c,q_85,usm_0.66_1.00_0.01,blur_3,enc_auto/a994932411404ef3bb797ba005125f5d.png)
![](https://static.wixstatic.com/media/a994932411404ef3bb797ba005125f5d.png/v1/fill/w_45,h_45,al_c,q_85,usm_0.66_1.00_0.01,blur_3,enc_auto/a994932411404ef3bb797ba005125f5d.png)
![](https://static.wixstatic.com/media/a994932411404ef3bb797ba005125f5d.png/v1/fill/w_45,h_45,al_c,q_85,usm_0.66_1.00_0.01,blur_3,enc_auto/a994932411404ef3bb797ba005125f5d.png)
bottom of page