Getting Started
On the home screen of your Site click the "Create Project" button to create a new Project.
Creating a Project will begin a 5-step workflow which will go through the process of importing data, setting data types, naming the project, and configuring relevant AI settings.
Step 1. Uploading Data/Choosing an Integration
The first step is to choose what type of data your Project will use - data from files you upload or data via an integration. This article will cover data uploaded. For integrations please see the dedicated integrations help article.
Click "Upload data from .csv or xlsx file" and choose a file. Your file will be uploaded and compared to a best practice checklist:
Field labels indicated by column headers
Column headers not empty
Sufficient text data (at least 250 rows)
Step 2. Select Column Data Types
Every column in your data will be mapped to a 'data type'. These are used by Kapiche to process the data in each column appropriately.
Kapiche automatically detects which data types might be best suited to each column, but it is best practice to check that they have been assigned correctly.
In addition, it is best practice to assign a unique ID field by clicking the 'SET UNIQUE ID FIELDS' button. Assigning a unique ID controls for any duplicate records. Having a unique ID will save you a lot of time when managing your data files for upload.
If you need any translation of text fields, choose the target language at this stage by clicking the 'LANGUAGES' button.
See below for a description of each data type that is available in Kapiche:
Text/Verbatim: Processed as text data to be used in the language model. Also used to display text excerpts/verbatims. Map your free-text/unstructured data to this data type.
Numerical: Processed as a numerical value. Numerical values can use mathematical functions in a query, such as "greater than" and "less than". Map any numerical structured data to this data type.
Date & Date Time: Processed as date formats to use as filters and segments. Also used to map data in timeline and trendline charts. Map any date fields in your data to this data type.
Category: Processed as a categorical value to be used as a segment. Compare across segments or filter your data and insights to a particular segment of the data.
List: Processes a list of values as categorical values, separated by commas. Each value in the list is counted toward segment counts.
NPS (0-10): Processed as a special value which is used to generate three categories (9-10 are labelled as Promoters, 7-8 are labelled as Passives, 0-6 are labelled as Detractors). Also used to calculate NPS and show NPS statistics. Map your respondent's recommendation score to this data type.
Ignore: Values in this column will be ignored and not included in this Project. Map anything you don't want to be included to this data type.
Once all columns are set to the correct data type, click Proceed.
Step 3. Project Details
At this step you have the option to enable AI themes, enable AI enrichments (if included in your subscription tier), opt into PII redaction, remove rows of data with dates in the incorrect format and configure the first day of the week.
Enable AI Themes
Enabling AI themes at project creation will allow you to provide context for the AI and the subsequent generation of AI topics at the Project level. You will have the choice at the Analysis stage to choose whether you generate AI Themes or not.
Enable AI Enrichments
Enabling AI enrichments will allow you to select which AI enrichments are generated on your Project data. Those that you select at Project creation will automatically be calculated and added to all subsequent data added to your Project.
PII Redaction
When this is enabled Kapiche will automatically detect personally identifiable information (PII) in your data, such as full names, phone numbers, addresses, etc, and replace it with a generic placeholder so that we are not storing sensitive information.
Note: While we make our best efforts to ensure maximum efficacy no solution in this space is 100% accurate.
Skip Bad or Missing Dates
This feature allows date fields with some invalid values to be successfully loaded and analysed. This is not enabled by default because often we find that failing on invalid dates provides us an opportunity to check if they can be remediated at the source. If not, then we can enable this setting in order to make the best use of the data we have.
First Day of Week
By default, Kapiche views a reporting week as Sunday to Monday. This option allows you to change which day a week should begin for the purposes of aggregation, default filtering and display.
Step 4. Dataset Configuration
Our AI will auto-suggest a dataset description based on the data you have uploaded at Project creation. You can edit this to provide the AI with the best context on the dataset and your organization.
Subsequent dataset configuration screens will ask you to confirm or complete the question that is being asked for each text field in your dataset.
In addition, you will be asked to select and detail any AI enrichments that you would like. For a full list of available AI enrichments see the dedicated help article available to subscribers.
See below for a description of some of our available AI Enrichments:
Emotion: Identifies the emotional state(s) expressed by the customer using a framework based on Plutchik's Wheel of Emotions (e.g., Joy, Anger, Gratitude, Confusion).
Journey Segments: Identifies which stage of the customer lifecycle the interaction relates to (e.g., onboarding, renewal, support), based on your predefined journey stages.
Product Feedback: Categorizes any product-related feedback mentioned by the customer into your predefined categories.
Custom enrichments: We can create a range of custom enrichments according to your specific needs and requirements.
Step 5. Finish & Create Project
Now that your Project is created, you can move onto the next phase, creating an Analysis!