by Alibek Utyubayev (alibek.ex at gmail.com) Analysis of factors that affect level of PM2.5 air pollution in the city of Almaty, Kazakhstan
In this analysis I will try to answer 3 following questions regarding air pollution in Almaty for the period of 2017-2020:
-
What factors can help us predict air pollution le vels and thus can be possibly affecting it?
-
How does air quality deteriorate during the colder weather and heating seasons?
-
How did city lock-down due to COVID-19, and decreased car traffic as a result of it, affect the air quality?
The structure of the analysis is little bit different and it is oulined in the Jupyter notebook. The analysis includes data extraction, training and selecting the best model and other data analysis related aspects.
In their City Development Strategy, Almaty city officials used to downplay the air pollution problem and the effect of heating pollution, while concentrating more on car traffic pollution. The situation has changed a bit after city activists created a petition to name air pollution as the city’s main challenge and demanded more actions in terms of decreasing pollution by power and heating plants. However, there still were some opinions attributing air quality improvement in April 2020 solely t o the decrease in traffic. Thus, it is especially interesting to see the answer to the third question. I lived in a city for 7 months now, starting from December 2019, and have witnessed noticeable air pollution. The COVID-19 lockdown provided me with a good opportunity to see the effect of a traffic that was reduced almost by 80% by some estimates.
The summary for general audience that does not go deeply into the technical parts is available at my Linkedin page: https://www.linkedin.com/pulse/analysis-factors-including-lock-down-due-covid-19-affect-utyubayev/
- analysis_pm25.ipynb - Jupyter notebook with the analysis
- daily_almaty_14apr20.xls - file with air pollution data from airkaz.org
- almaty_weather_201701_to_202006.xls - file with weather data from rp5.ru
SHAP Seaborn Matplotlib Pandas Numpy Sklearn Xgboost Copy Math
Used a function from ChrisKuoColumbiaU's GitHub repository for SHAPLEY values visualization