All Collections
Getting Started with Kapiche
Part 2: Analyzing your first dataset
How to use and get the most out of your Kapiche Queries
How to use and get the most out of your Kapiche Queries

Follow these tips to get the most out of your Kapiche Queries!

Josh Winters avatar
Written by Josh Winters
Updated over a week ago

Queries are one of the most powerful features Kapiche has to offer. With queries, you can find almost anything you want to see from your data and save them as themes.

Below, you'll find our video and written guides that walk you through everything you need to know about how to use and get the most out of your Kapiche queries!

Video walk-through 🎥

Keep scrolling for our written guide instead 📓 👇

What are Queries?

Queries are an advanced form of search on your data. They are made up of one or more "lines" that each impact the results of the query. You can think of each line as a filter, with the goal of fine-tuning the results to however broad or specific you want them to be.

To start building a query, click the search bar along the top:

From there, you can use four key elements to build your queries:

  • Fields - Filters results using the different structured data fields in your data set.

  • Concepts - Filter results using words found in a verbatim. The top concepts in your dataset will be coloured. All words that are queried in Kapiche will match variations of the word -- for example, querying product will also match products, Products, Product and so on.

  • Terms & Phrases - Filter results by manually typing a specific word ("service"), phrase ("customer service") or even a sentence ("excellent service was provided")

  • Explicit matches - Filter results using the exact word or phrase by wrapping the search terms in quotation marks. For example, querying "OVERDUE" will only return verbatims that have have explicitly written OVERDUE. This works for phrases too, in which the phrase must be written in the same order and capitilisation as you search for.

As well as query operators that further fine tune your results:

  • OR operators - Each line in a query can include multiple elements of the same type joined by an "OR" operator. Using an "OR" operator means the query results must adhere to at least one of the linked elements.

  • AND operators - Using an "AND" operator means the query results must adhere to the elements of all query lines (more on this in the example below).

  • DOES include / DOES NOT include - Allows you to determine whether you'd like the elements of each query line to be included or excluded from your query.

  • Date Ranges - If your data includes date ranges you can use the date filters to filters results that match a specific day or match time frame parameters (such as "between Date X and Date X" or "Before/After Date X"

  • Numerical Operators - If your data includes numerical fields (such as CSAT scores, revenue buckets, etc.) you can use Numerical Operators to filter based on things like "Revenue is less than $X" or "Revenue is between $X & $Y".

To help demonstrate, let's look at a sample query for the concept 'Food':

What we’re looking at now is the results of all the passengers who mentioned “food” when leaving feedback about their experience.

But let’s say we want to combine the results of the passengers who mentioned “food” with those who mentioned “drinks”. What we can do is click the “OR” button next to “food” and then type in “drinks”.

Kapiche is now showing us the results of the passengers who mention “food” OR who mention “drinks” (they don't need to mention both terms).

Alternatively, if we want Kapiche to show us the results where passengers mention both “food” AND “drinks” in the same response, we’ll need to use the AND operator.

To do this, we'll remove drinks from the first query line by clicking on it and choosing "remove from line". From there, we need to create a second query line for "drinks" by

typing “drinks” into the search bar towards the top of the page and hitting enter.

Because we’ve added “Drinks” to the query on its own line, our query is now showing us the results for the responses where passengers mention both “food” and “drinks” in their verbatim.

We can also switch between "and includes" to "does not include" (e.g. if we'd like to see responses where ONLY "food", specifically excluding verbatims where "drinks" was mentioned).


To do this, click on the "and includes" button and change that to "does not include":

Comparing Queries

One of our most-loved query tools is the ability to run comparative queries side by side, as this makes it super easy for you to compare how different customer segments, like regions, revenue categories and demographics are performing vs your overall company average.

For example, if we run a comparative query between those who flew Premium Economy vs those who flew Standard Economy, we can immediately see how the respective NPS and narrative changes between the two customer segments:

NPS & Sentiment Comparison

Narrative Comparison

Adding Segments to a Query

You also have the option to add Segment filters to your query using the structured fields included in your data set by clicking on the query search bar along the top and choosing the field you would like to filter by.

From there, you will need to choose the relevant field segment:

By adding in the ‘NPS: Detractor element, we’re now looking at the results of the passengers who mentioned “food” AND "drinks" in their response, who are also a Detractor.

Adding Date Ranges to a Query

If your data set includes date ranges, you can also run queries with specific date filters, such as “Show me results of the passengers who flew prior to Nov 6 2019”:

Using Data Visualizations to Add to a Query

You'll also find a variety of data visualizations throughout your query page which can be used to quickly add new lines to your query:

  • Clicking terms on the Context Network or Context Graph

  • Clicking a label on the sentiment bar

  • Clicking a label on the NPS panel

  • Clicking a value on a Segment or Concept correlation chart

Saving Queries as Themes

To save a query as a Theme, simply create a query and click the "Save As..." button:

You'll then see the following dialog where you can define a name for the theme and choose to auto-add it to your Dashboards:

To access your saved Themes, simply click the "Themes" dropdown. Clicking a Theme will clear your current query and input the selected theme's saved query automatically.

Updating a Theme

Whilst a Theme is selected, any changes you make to the query can be saved using the green "Save" button. You can also simply save it as a new Theme.

Renaming & Deleting Themes

You can rename and delete Themes by clicking the respective icon in the Themes dropdown menu in the Queries screen.

Using Themes on Dashboards

An important aspect of Themes is the relationship they have with the Dashboard. Dashboards require Themes to display information. Dashboards display various visualizations and overviews using the data from each Theme selected to be shown on the Dashboard.

When you save a query as a theme, it becomes an option to be shown on the Dashboard. Only themes that have a Concept, Term or Phrase element (unstructured data) will be available, since the Dashboard has its own separate structured data filters.

Questions? 🤔

If you have any questions about creating queries (or anything else!) you can get in touch with us any time by hitting the blue chat button to your right 👉


Did this answer your question?