Bekeley Logo

CalSHARE Data Hub

How to...AboutTechnical Details
Take me to the dataBanner

Technical Details

Percentile rankings: Percentile rankings included in the indicator charts were calculated by arranging in order all unique values for a given indicator across California, identifying the position of that geography within that ordered set of values, and dividing that position by the total number of unique values (i.e., # rank/total unique values). Often, the direction of this ranking needed to be flipped so that an indicator that demonstrates an undesirable health characteristic (e.g., homelessness) displays a percentile that is low if that undesirable characteristic is high and a high percentile if that undesirable characteristic is low for the selected region, for example. This was done by subtracting the percentile calculation described above by 1.

Imprecision: Some estimates included in the dashboard are imprecise or "unstable". These are denoted by an asterisk next to the indicator value. An unstable estimate means that there is too little data for that geographic region to be certain that this measurement is accurate. These estimates are still valuable as a marker of an approximate value for this region but should not be used to make programmatic or policy decisions.

Data limitations:

Regions with no data. Many of the data sources include some geographic regions for which there is no data. This is often due to the area having a small population. Regions on the map for a specific indicator that are greyed-out indicate no data is available.

Race/ethnicity and gender subgroups. When the data source allows, indicators can be broken down by race/ethnicity and gender. We want to acknowledge that the definitions and categories for these groups are based entirely on those which are used and available from the data source. Nearly all data sources do not include nonbinary gender categories and do not differentiate between sex assigned at birth and gender; there is still a great need for improving measurement of gender in census and surveillance data. Similarly, many of the data sources use broad race/ethnicity categories that often align with those used in the census; more nuanced data on race and ethnicity is needed.