Visualization on Beijing Air Quality

Result

  • You can try it here.

Data Description

  • Time range: 2012.10.17 ~ 2012.10.30, per hour
  • Places: 26 monitor stations (1 place missing)
  • Attributes: CO and O3 values in Beijing City; SO2, NO2, PM10 and PM2.5 values in each monitor stations.

Design

My design aims at:

  1. Distributions: Air quality distributions in place and time.
  2. Comparisons: Compare among regions, and among locations can be obtained.
  3. Surveillance: Both the abnormal values and the missing data can be detected easily.

About the Interface:

  1. Line graph: 6 lines for 6 attributes of the whole Beijing City changing in 14days. That means except the first 2 values which are already given, the last 4 values are average value of all existing places.
  2. Block Matrix, each block is an icon consists of 4 part, for express one place's status in on hour. The whole matrix represents all 26 places' air quality in a whole day, unfilled blocks means missing data of related attributes. Clicking on the line chart above will decide which day to show.
  3. Tree-like structure of places, since Beijing City consists of 5 large regions, I labeled each places

Implement

Data Pre-processing

  1. Remove the missing place. After explore the whole data, I found the 14th place doesn't appeared at all, so I just remove the 14th place from the place list, that means 26 places left now.
  2. Labeling the places into 5 large regions.

Using D3

  1. Draw those graphs and block matrix mentioned before.
  2. Add linking between line graph and block matrix, and labels on hover, etc.

Results and Findings

  1. Access: You can try it here.
  2. Distributions: The worst time appeared at night of 26 Oct in 14 days. And seems like bad air usually appeared in afternoon or at night.
  3. Comparisons: Most likely polluted places are those in 西北 regions.
  4. Surveillance: We can see that 石景山古城's records disappeared since 10.19 16:00, so maybe there's something wrong with the monitoring system.
  5. Abnormal Peak discovered easily of SO2 value in 顺义新城 at Oct 22, 15:00.

To-Improves

  1. Space compression :p
  2. Map of Beijing needed for clearer perception in space. Maybe an abstract or topology one is enough.
  3. Instructions and explanations on how to explore.
 
public_course/visclass_f12/assignment/a02/qingya.shu/start.txt · Last modified: 2012/12/03 02:19 by qingya.shu