Analysis of wildfires in California


Visualization

The Backstory

There’s been more and more discussion of pervasive wildfires in California in the news. Much of which has been linked to climate change. This spawned an interest to see how much the occurrences have increased over the last decade. After a thorough search of available datasets, I managed, with a little help, to find data dating back to 2010 to illustrate the salience of the growing wildfires. (Photo Credit: Matt Howard).

Process and Tooling

Datasets:

Tools:

  • Carto: used to help synthesize and analyze spatial data.

In order to easily upload the data volumes ( the 2010s data in particular) I found that the Shapefile worked best in Carto.

Once in Carto, I used two layers for the analysis. One that offered a view of the data that was current 30 days prior to July 13th, 2021 to give a view of the middle of the summer. The other offered data from 2019 back to 1950s, but only had the labeled data back to 2010. This offered the ability to analyze how the wildfires in California were progressing over the course of 10 years.

I also found that in order to put the focus on a single state, I needed to apply a filter on the state column. When the filter was applied, I quickly determined that I had to change the label of the column to reflect a more descriptive filter title.

I also found that I had to change or remove the names of the datasets to make the legend more legible.

The legend labels were easily modified by using the “Creating your legend” option. The only caveat here is that I had to modify the legend colors separately from the colors of your data points. When you have to make 13 color selections, this is a bit tedious.

Due to limitations, I had to use two color tools to create a 13-color palette: colorbrewer2.org and color.adobe.com/create/color-wheel. To determine the best color layering option, this resource was particularly helpful. I wanted to be able to easily distinguish the wildfires within the context of the year they occurred.

I opted for the src-over blending option because it offered full transparency of the source and destination layers, allowing the reader to see all layers involved in the overlap area, which is easier to see when zoomed in.

The color choice was based on a couple of variables. I wanted to create a sense of urgency, or rather, light a fire, over the concern of increased forest fires that have increased not only in number, but in size. I felt this using the warm color schema would accomplish this when juxtaposed to the “Dark Matter” base map.

  • Column name change
  • Output for filter name
Changed (A) and removed (B) dataset names

Legend options and modifications

Style selections for Polygons

I left the city labels visible to help the spatial components resonate more with the reader. Then, I added in a clickable pop-up in the dark theme to match the background that illustrated the year and size of the fire in acres, as measured by a GIS system for better accuracy to ensure that the reader could assess the year and magnitude of the wildfire’s impact by location/polygon of choice.

Results

Data output from Carto

The combination of colors sets the right tone for the reader to immediately grasp the issue. The dark background almost gives the sense of the land turning to ashes. The clickable pop-ups along with the zooming features also do a good job of offering more context and interaction for the reader to engage, and explore to get a greater sense the data.

Leaving the city names also gives a better sense of location. For people who live in the United States, the larger cities that are more well known such as Sacramento, San Jose, or San Fransisco, could help the visual resonate more with the reader.

Src-over blending shows the layered effect even when zoomed out; making it an effective tool to highlight the past year’s data. Especially since the user can click to get the precise date and surface area of each wildfire. However, even with the src-over blending effect, apart from clicking on each polygon’s pop-up, it may be difficult to determine which fire happened in each year.

At a minimum, the color choice is appealing to the eye and there’s a good contrast with the black background to easily determine where the wildfires are on the spatial visualization.

Next Steps

Given more time, I would have invested more time on the color palette to make the years appear more distinct. The blending choice did reveal layers well, but the colors still blended too much to easily determine which year a wildfire occurred. I also would have done more digging to see how to apply a measurement scale to give an initial idea of the acres without having to click on a pop-up for the reader. But in my initial assessment, I couldn’t figure out a way to apply this feature. It may be a limitation of Carto.

References

  1. 2010s—California Open Data. (n.d.). Retrieved July 13, 2021, from https://data.ca.gov/dataset/2010s3
  2. CartoCSS Compositing Operations. Retrieved July 13, 2021, from https://carto.com/help/tutorials/cartocss-compositing-operations/ KPIX CBS SF Bay Area.
  3. Cal Fire Says Fires This Year Have Already Surpassed Record Destruction of 2020 Season. Retrieved July 19, 2021, from https://www.youtube.com/watch?v=g0VGwDS6fck KPIX CBS SF Bay Area.
  4. Northern California Wildfires Rage Uncontained. Retrieved July 19, 2021, from https://www.youtube.com/watch?v=OGKt6yU1EWk
  5. The science of how climate change impacts fires in the West. (2020, September 17). Science. https://www.nationalgeographic.com/science/article/climate-change-increases-risk-fires-western-us
  6. Wildfire Perimeters—California Open Data. (n.d.). Retrieved July 13, 2021, from https://data.ca.gov/dataset/wildfire-perimeters1

No related posts.