LOS Angeles COVID-19 MAPs

The map and linked dashboard (below ) contains visualizations of data pulled from the Los Angeles Times’ Covid-19 GitHub for “places” Los Angeles County and a research article (Yearbook of the Association of Pacific Coast Geographers – 2021 )

Map 1:  Infections per 100,000 persons as of March 15, 2021   Mapped at the 350 “micro neighborhoods” level during the period of very low rates following the largest spike in infections during the winter of 2020-2021.  Weekly rates at this time were both modest and consistent across the county.

Map of Infection Rates by All Neighborhoods per 100,000 persons as of March 15, 2021.  Lower rates are evident in the coastal neighborhoods and foothill/mountain regions.  Higher rates are in the Northeast San Fernando Valley, East and South Los Angeles.

Map of Infection Rates by All Neighborhoods per 100,000 persons as of March 15, 2021

 

Map 2: Map of COVID-19 Infection Rates consolidated into 66 regional neighborhoods because of troubling spatial autocorrelation among model residuals.  Note the vast difference in infection rates from darkest blue (low) to red (high) regions on the map.

COVID-19 Infection Rates (March 15, 2021) by Super Neighborhood.  Lowest rates in the coastal regions and the highest rates were in the northeastern San Fernando Valley as well as neighborhoods east and south of Downtown Los Angeles.
COVID-19 Infection Rates (March 15, 2021) by Super Neighborhood.

 

Map 3:  Model Residuals Map  The final model “predicted” the COVID-19 infection rate extremely well (R2 = 0.93) using housing variables (Average Household Size ** , Percent Renter**,, Percent in Group Housing) as well as Income and Ethnicity (Percent Hispanic  and Percent Asian).    The neighborhoods in red had higher than predicted rates and the neighborhoods in blue had lower than predicted rates.  Both indicate the possibility of another lurking variable (like air pollution?) that may help identify a final cause of variation in the infection rates.

Los Angeles County
Map of Model Residuals showing regions where the model underpredicted (red) or overpredicted (blue) the rate of infections in LA County by neighborhood during the first year of the pandemic.
Ordinary Least Squares Regression Residual Map. Model was well specified and robust (R square = .93)

Community and Trend Analyses

The three-page dashboard below contains a basemap of cumulative cases and infection rates for most communities or neighborhoods in Los Angeles County, along with interactive graphs that permit users to view recent ‘curve’ trends by community and comparative ‘curves’ for multiple communities.  A third map shows the contribution of nursing homes and residential care facilities to the overall caseloads by community.  

Full-Screen Dashboard 

Southern California – Residential Care/Nursing Homes

Please check out my friend Lauren He’s map of Covid-19 cases in Nursing Homes and Residential Care Facilities.  It’s a very important angle on the disease that is both under-reported in the press and for which it is difficult to obtain useful data.  Bonus:  Lauren’s a high school senior from Austin Texas.  

https://arcg.is/1CveCC

 

9 thoughts on “LOS Angeles COVID-19 MAPs”

  1. I don’t understand what the color dots mean can you please explain what the yellow and green are I know the red means hot spot.

  2. Thank you for such a detailed site! We live in Tucson and will be driving to Inglewood, CA. soon to attend a funeral. It’s nice to see that the spike in Covid cases appears to have subsided in that local area as opposed to only seeing L.A. County’s Covid surge numbers. This information was very helpful in determining our course of direction etc.. We will be wearing our masks and social distancing too, of course, as our area is also just slowly growing by single digits.

  3. Hi! Thank you for a great interactive map! Have a question: I’m monitoring the emergence of covid cases in the Melrose area and notice that on 7/6 2020 the new cases are a whooping 928 from a regular in the low double digits, indeed they equal the cumulative cases for that very date and I wonder if this perchance is a typo/data error. Thank you so much!

    Christina

    1. It’s a data error associated with the lack of data over the July 4th Holiday. My inability to get the software to ignore days with zero information is the problem…so the algorithm assumes that there were zero cumulative cases on those dates, and then it appears as a big jump in cases when the data feed resumes.

      1. Steve, wouldn’t it be better to find dates with missing data then, and simply carry over the previous days data to avoid these issues. I’d be more than happy to provide you with some data cleanup scripts if you would like. Also, have you thought about changing the colors to indicate where things are growing most rapidly (averaged values over X days) instead of total cases. I found this page trying to find a heat map of where the spread is slowing down and where things are still out of control.

        1. Kyle – there’s a LOT of things I could do, but I’m up to my eyeballs in other things. If you want to eliminate missing dates, it’s easy to click on them and chose “exclude”. This is mostly proof of concept with a software that is experimental for me.

          There are various heat maps (nationally) using a variety of heatmaps. I provide a couple of links on this page. The ESRI maps are very good in that respect.

      2. Aaargh! Frustrating, I imagine.
        MANY thanks for all your diligence and problem-solving!
        We non-mathy mortals quail at the thought!

Comments are closed.