Search queries spanning all of 2010, including both keyword searches and any search filters applied. Data files: Jan - June 2010 ~6MB zip, ~42MB CSV, 531K records, ~2.3M searches July - Sep 2010 ~5MB zip, ~36MB CSV, 448K records, ~1.7M searches Oct - Dec 2010 ~8MB zip, ~56MB CSV, 665K records, ~2.8M searches The search data files are exports from Google Analytics. As such, each search record appears as a URL with parameters representing any keyword or filter-based search criteria. Each URL also has an accompanying count that indicates how many times it was executed during the period included in the export. Since a Google Analytics provides a few additional metrics in its standard data export, we're including them too. Because we were unable to export all of 2010 at once from Google Analytics, the searches are spread over 3 files. Since each of these files is a roll-up of the searches conducted during its respective time period, there are duplicate entries across the files and you would need to merge the counts for each search across the files to arrive at the annual counts. Unfortunately, we're not able to provide any user-session information so you cannot join the searches executed by the same user in a given session. However, you can infer the order in which multiple filters are combined as they appear left-to-right in the URL, eg. the first search parameter in the URL was the first filter the user applied. Page: the search, including both keyword searches and any search filters applied. More details below. To see a list of all the search filters available to prospective donors, look at the right column of our search page and choose "More Criteria" to show the full list. You'll see the ways prospective donors can search for projects, such as location, subject, keyword, resource type, grade level, whether the project is eligible for any match offers, whether the project already has donations, etc, etc. Since our front-end search URLs and JSON API request parameters are nearly identical, the JSON project listings request documentation defines all the search URL parameters you'll encounter. To get a feel for the mapping, you can simply paste any search URL from the log into our front-end and you'll see the search parameters appear in the breadcrumb navigation. For example, pasting the URL "/donors/search.html?state=MA&proposalType=2&schoolType=1" into our front-end shows "Project requests > Massachusetts > Technology > Charter" in the breadcrumb navigation. Also note that the URL "/donors/search.html?schoolType=1&proposalType=2&state=MA" shows "Charter > Technology > Massachusetts" in the breadcrumb navigation: it shows the same classroom project listings, but reflects a user who added the same search filters in a different order. If you see a parameter that's not in the JSON API or explicitly listed below as OK to ignore, you can ignore it but we'd encourage you to ask about it via the discussion group. Add'l notes about the search URL parameters that are covered by the JSON project listings request documentation: max: Our UX defaults to 4 projects per page pagination and enables users to increase their pagination to 10, 25, or 50. If you see other pagination amounts, they are the result of users hand-editing the front-end URL. Our UX also remembers users' pagination preferences in a cookie which results in many non-default pagination requests from our non-new users. keywords: Due to internal site search tracking in Google Analytics, the "keywords" parameter is repeated twice in the search URLs, coupled with the keyword search strings. For example, a keyword search for "environment" will display as "/donors/search.html?keywords+environment&keywords=environment", with the first displayed keywords+environment being added by Google Analytics. This neither has an impact on the destination URL, nor was it displayed as such on the front-end during the original search. These search URL parameters are not covered by the JSON project listings request documentation:
These search URL parameters can be ignored, as they do not reflect any particular user intent nor do they alter the search results:
 This work by DonorsChoose.org is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported License (CC BY-NC 3.0). If you'd like to use this data for commercial purposes, drop a note to apiquestions (at) donorschoose {dot} org and tell us a bit about your plans. Our strong preference is to greenlight your commercial application with no licensing fees, and we have never charged for access to our API or data. We just need to make sure that the application won't run contrary to our org's mission or abuse the rich content that our teachers have created. |
|
Open Datasets >
Search Data |