Creating a custom filter is a great way to reduce irrelevant information from appearing in your saved searches.
When creating a new filter, you may choose between a Keyword Inclusion, Keyword Exclusion, Author Exclusion, or Object Detection.
Any posts returned from this filter must include at least one of the specified keywords.
Example: Below is a keyword inclusion filter to narrow down results from a Notre Dame location search. The results from this filter must include the word "fire" or "burn" in English (and their translations in French).
Any posts returned from this filter will NOT include any of the specified keywords.
Example: Let's say your job was to evaluate the public sentiment for the camera brand, Canon. You create a saved search with "Canon" as the keyword, and apply both positive and negative sentiment filters to it (please see these filters within our pre-built filters section). You find a lot of your posts returned are actually regarding people speaking about cannons, but misspelling it as "canon".
By creating and applying a keyword exclusion filter with keywords related to cannons, you could weed out this unwanted information from your returned posts.
Results must NOT be posted by the specified authors.
Example: Let's say your job is to monitor the Microsoft Headquarters in Redmond, WA, so you create a geo-fence around the location and save it into your saved searches.
As you can see, some of the results are from job posting boards within the area. To ensure these don't appear within your search results, you could apply an author exclusion filter with the handles of known job-posters.
Posts must include specified objects.
Example: The Peace Arch Border Crossing is the 3rd most popular Canadian/United States border crossing. If your job was to monitor this area, you might draw a geo-fence around the location.
To narrow down the search, you could apply an object detection filter to filter for specific objects of interest, such as pictures of passports. Our object detection currently is able to detect over 20,000 objects from our supported keywords.
Note that when creating inclusion or exclusion search filters, any keywords you input will also include the variants of the root word.
This means that entering "burn" into your keyword inclusion filter will also aggregate result containing “burns” and “burning”, since “burn” is considered a prefix of those words.
The inverse is also true- this means that if entered “burn” as a keyword in an exclusion filter, Echosec will also remove words containing your keyword as a prefix (“burns”, “burning”), but not every instance of this combination of letters. For example, “auburn” still could be included in results.