The group members

Zhuolun Huang, zh2494
Fei Sun, fs2757
Lin Yang, ly2565
Yihan Qiu, yq2321
Weiheng Zhang, wz2590

The tentative project title

The Impact of COVID-19 Lockdowns on Air Quality in Chinese Cities

The motivation for this project

In order to contain the COVID-19 outbreak, China entered lockdown in early February 2020, minimizing all industrial, transportation, and commercial activities. Although the lockdown caused tremendous loss to the economy, previous research has shown the air condition in the top four megacities of China improved significantly during the lockdown period, due to a dramatic decrease in automobile and industrial emissions. Our motivation is to find out whether “lockdown improves air quality” is a common phenomenon in Chinese cities including but not only limited to megacities. We are also interested in how the geographical location, population, and GDP of each city affect the degree of air quality improvement, and which pollutant’s daily air quality index(AQI) was affected the most. AQI is a variable describing the daily level of air pollution from pollutants such as PM2.5, PM10, NO2. Each pollutant has its own AQI value for each day.

The intended final products

We will present interactive graphs and thematic maps showing the degree of improvement of daily average AQI for multiple pollutants, among thirty representative Chinese cities during lockdown compared to air quality data in the same months of the year 2019.
We will also build interactive graphs showing how the population and GDP of cities in the same geographical district correlate to the degree of improvement of daily average AQI.

The planned analyses / visualizations

We will make various plots to evaluate how lockdown policies affected air quality in most Chinese cities, and how population and GDP affected air quality improvements in these cities. We will also analyze which cities had the most and the least air quality improvements and figure out if there is an association between air quality improvements and geographical locations. Here’re our expected results:

  • A bar graph of mean AQI difference for each air pollutant(PM2.5, PM10, SO2, O3, NO2, CO) between 2019 and 2020.
  • A boxplot of daily PM2.5 AQI from Feb. to Apr. (lockdown period) in the past year (2019, 2020) in 10 representative cities.
  • A bar graph of mean PM2.5 AQI from Feb to Apr in the past three years in 10 representative cities.
  • A line plot of trends in monthly average PM2.5 AQIs from 2019/02 - 2020/06 in 10 representative cities.
  • Scatterplots of PM2.5 AQI differences vs. GDP and PM2.5 AQI differences vs. population in all cities.
  • A map of China showing each city’s AQI difference between early 2020 and 2019.

Coding challenges

  1. Although we already have accurate data on air quality in each city, the data of every 30 cities are independent files, so it is tedious to combine the air pollution data of each city into one file. Comparing data from different cities could be complicated.
  2. To evaluate the relationship between population/GDP and air quality improvements, we have to manually gather data from the web.
  3. There are some errors and obscure information in the air quality data, so we need to review and investigate the data of each city carefully before we proceed.
  4. It can be time-consuming to generate reader-friendly plots, especially plots containing multiple cities.

Timeline

Date Work Assignment due
11/13 Complete proposal 11/13, 1pm
11/16-11/19 Project review meeting with TA 11/16-11/19
11/20-11/21 Complete data collection and division of labor NA
11/21-11/26 Individual data analysis NA
11/27 Group meeting: discuss everyone’s work and any problems NA
11/28-12/3 Individual work NA
12/4-12/7 Webpage design and screencast 12/11, 4PM
12/7-12/10 Group meeting: finalize report NA
12/11 Complete report 12/11, 4PM
12/11 Peer assessment 12/11, 8PM

The intended final products

We will present interactive graphs and thematic maps showing the degree of improvement of daily average AQI for multiple pollutants, among thirty representative Chinese cities during lockdown compared to air quality data in the same months of the year 2019. We will also build interactive graphs showing how the population and GDP of cities in the same geographical district correlate to the degree of improvement of daily average AQI.

Coding challenges

  1. Although we already have accurate data on air quality in each city, the data of every 30 cities are independent files, so it is tedious to combine the air pollution data of each city into one file. Comparing data from different cities could be complicated.
  2. To evaluate the relationship between population/GDP and air quality improvements, we have to manually gather data from the web.
  3. There are some errors and obscure information in the air quality data, so we need to review and investigate the data of each city carefully before we proceed.
  4. It can be time-consuming to generate reader-friendly plots, especially plots containing multiple cities.