Collecting and Cleansing AQI Data
Collect and cleanse hourly AQI of five cities of China （where U.S. embassy locations: Beijing, Shanghai, Guangzhou, Shenyang, Chengdu) from 2008 to 2014.
Link to raw and cleansed dataset：Baidu Cloud.
- Data Source: stateair.net
- Create Date: 20150406
- Updata Date: 20150407
Defination of the dataset and introduction can be downloaded from this link: (http://www.stateair.net/web/historical/1/1.html)
Data use statement (http://www.stateair.net/web/historical/1/1.html)
Files:
- City_PM25_From_To.csv
CSV dataset of cleansed data of city - USEmbassyAQI.csv
CSV dataset with five cities combied - USEmbassyAQI.RData
In Rdata format - Raw/*
Raw dataset
Cleansing Process
- Delete the first three rows of text.
- Delete the “Parameter” column, since all have the same value “PM2.5”.
- Delete “DATE(LST)” column, keep “Year”, “Month”, “Day” and “Hour” columns.
- Delete “Unit” column, since all have the same value “ug/m3”.
- Delete “QC Name”, since all are blank.
- Transfer “Duration” column, which takes two values of “Valid” and “Missing” to a boolean column “Valid”.
Update 20150407
There are 17 pairs of duplicates (same location and time), all of which are at 3:00 am. The duplicates have different values of AQI and the 2:00 am records are missing for all the cases, so correct the first appearance of 3:00 am as 2:00 am record. The following data is adjusted:
| Row | Site | Year | Month | Day | Hour | Value | Valid |
|---|---|---|---|---|---|---|---|
| 6675 | Beijing | 2009 | 3 | 8 | 3 | 179 | TRUE |
| 15579 | Beijing | 2010 | 3 | 14 | 3 | 75 | TRUE |
| 24315 | Beijing | 2011 | 3 | 13 | 3 | 351 | TRUE |
| 50523 | Beijing | 2014 | 3 | 9 | 3 | 55 | TRUE |
| 59355 | Chengdu | 2012 | 3 | 11 | 3 | -999 | FALSE |
| 68091 | Chengdu | 2013 | 3 | 10 | 3 | 176 | TRUE |
| 76827 | Chengdu | 2014 | 3 | 9 | 3 | 65 | TRUE |
| 85683 | Guangzhou | 2011 | 3 | 13 | 3 | -999 | FALSE |
| 94419 | Guangzhou | 2012 | 3 | 11 | 3 | -999 | FALSE |
| 103155 | Guangzhou | 2013 | 3 | 10 | 3 | 92 | TRUE |
| 111891 | Guangzhou | 2014 | 3 | 9 | 3 | 23 | TRUE |
| 120747 | Shanghai | 2011 | 3 | 13 | 3 | -999 | FALSE |
| 129483 | Shanghai | 2012 | 3 | 11 | 3 | 89 | TRUE |
| 138219 | Shanghai | 2013 | 3 | 10 | 3 | 107 | TRUE |
| 146955 | Shanghai | 2014 | 3 | 9 | 3 | 55 | TRUE |
| 155739 | Shenyang | 2013 | 3 | 10 | 3 | -999 | FALSE |
| 164475 | Shenyang | 2014 | 3 | 9 | 3 | 53 | TRUE |
Moreover, there are records with value -999 but not marked as invalid in Shanghai, such as:
| Row | Site | Year | Month | Day | Hour | Value | Valid |
|---|---|---|---|---|---|---|---|
| 127905 | Shanghai | 2012 | 1 | 5 | 9 | 49 | TRUE |
| 127906 | Shanghai | 2012 | 1 | 5 | 10 | 49 | TRUE |
| 127907 | Shanghai | 2012 | 1 | 5 | 11 | -999 | TRUE |
| 127908 | Shanghai | 2012 | 1 | 5 | 12 | -999 | TRUE |
| 127909 | Shanghai | 2012 | 1 | 5 | 13 | 50 | TRUE |
| 127910 | Shanghai | 2012 | 1 | 5 | 14 | 50 | TRUE |
The two records in the middle is invalid but not properly marked. There are totally 296 cases like this that are adjusted by marking “Valid=FALSE”.
Wang, Guansong
Data Worker
My research interests include distributed robotics, mobile computing and programmable matter.
