7 Using Advanced Options
How do I use the filters on my data?
This feature is only available in the website tool
You can use the filters to analyze a subset of the points in the dataset that you upload to the website tool. The tool allows the following filter types:
Text: Filter by text values in a selected column that are equal to or not equal to one or more values. Multiple values can be entered separated by commas. In such a case, the values will be evaluated as “or” conditions for which rows will be kept if the selected column equals or doesn’t equal any selected values.
Numeric: Filter numeric columns by equal, not equal, less than, less than or equal to, greater than, or greater than or equal to a number. You can add multiple numeric filters to the same column and filter multiple columns.
Date: Filter to keep rows with a particular date or date range.
If you set multiple filter conditions, the website tool will only use rows that meet all conditions. For example, if you select two numeric filters and two date filters, only rows that match all four filtering conditions will be returned. This means that certain filtering operations, such as filtering to data in either of two date ranges, are not possible.
If you want to filter your data in a way that is not enabled by the website tool (such as regular expressions or geographic filters), then you need to filter your data before uploading. The tool detects whether your column is text, numeric, or date on the basis of the first 10 rows in your dataset. The tool will only recognize a column as a date column if it follows the ECMAScript date time string format (for example, YYYY-MM-DD). Columns with just years (for example, 2014) may be recognized as text columns. For help understanding how our tool recognizes column types, please see this page.
Users should perform any filtering on their data before submitting to the API. Unlike the web tool, the API does not allow for users to filter records in the resource dataset after submission.
How does changing the weights affect my results?
The weights determine how each point is counted when measuring representativeness. If your dataset is bike share stations and you select to weight by number of bikes, then the tool will weight a bike share station with 10 bikes 10 times as much as a bike share station with 1 bike when constructing geographic and demographic disparity measures. If you do not select a weight, then each row (i.e., geographic point) in the data is weighted equally. Rows with a weight of 0 are treated as null/NA values and discarded from the analysis. Bear in mind that weighting your data will affect both the geographic disparity score shown in the map and the demographic disparity scores shown on the chart.