Algorithm - A series of steps followed by a computer program in order to produce a desired output. A lot of computer science research focuses on finding the simplest or fastest algorithms to perform certain tasks. Keatext’s machine learning algorithm is specifically trained to extract customer feedback, finding the relevant Topic and its Opinion companion.
All Related Comments bar - Clicking this bar at the bottom of the workspace pulls up a panel of Comments related to the selected Comment – you can increase the quantity from 6 to 12. Each Comment is clickable through to the original Record.
Analytics page - The bulk of results are displayed here.
Artificial Intelligence (AI) - Refers to computer programs which are capable of performing complex tasks usually done by humans, such as recognizing faces in an image or, as in the case of Keatext, identifying customer opinions in text.
Blacklist - Contains results that have been hidden. This feature is used to clean data of unwanted noise.
Collection Date - A timestamp (usually when feedback was originally collected). If set during the upload, Time filters will be enabled.
Column Headers - These are the titles of the columns in the CSV file. These will be displayed in the Record filters. CSV files should be structured into clearly marked column headers.
Comment - A Comment is each statement in the record that contains a Topic and an Opinion, and has a Sentiment assigned to it. A Record can contain numerous Comments and the same Comment can be present in numerous Records.
Comment Summary - These can be accessed by clicking on any comment in the list. It contains the number of Total Commenters, the Sentiment Score, the Level of Interest, Correlations and Related Comments.
Commenter - The person who is giving feedback.
Commenter ID - A unique customer/employee identifier, such as an email or internal customer/employee ID linking an individual across multiple channels.
Comments Overview - Contains the number of Commenters in the selected sources. If you do not set a Commenter ID, the Comments Overview will be blank. This is useful for understanding the ratio of Comments to Commenter.
CSV - A file format used to store tabular data, such as a spreadsheet or database. CSV is the only file type that can be uploaded onto our platform. For analysis purposes, a typical CSV file should be structured into clearly marked column headers with at least one column containing customer comments and, if available, metadata.
Dashboard - This is where you will find a visualization of your data (currently not available to all users).
Data Source - This is the CSV file you upload to the Keytext system. You can upload multiple Data Sources to be analyzed separately or together, in any combination you choose. All of your uploaded Data Sources are accessible on the Sources page of the platform.
Data Type - This is the classification that specifies which type of value a field contains in the original CSV file. The options are: Boolean (logic such a true/false and yes/no), string (general characterization, such as language), number (a price or any other numeric value) and datetime.
Date Format - The different date and time representations supported by the system.
Delimiter - A character that marks the beginning or end of a unit of data. The Delimiters supported by the Keatext system are a comma, semicolon, or tab. The Delimiter will normally be auto-detected when you upload a CSV file.
Exports - All exports are downloaded as a compressed file containing the requested export in CSV format. If the results exported come from different datasets, each one will be a separate file.
Filterset - Filter combinations that can be saved for later use. A Filterset is composed of one or more filters. They can be as simple as a single Time Filter or as complex as multiple Record Filters used in conjunction with other filter types.
Groups - A variety of words and phrases that express the same idea grouped together by Keatext’s AI technology. You can manually modify groups if you wish.
Hide - This sends the selected results that have been checkmarked to the Blacklist.
Hide All - This sends all of the results to the Blacklist.
Interest - This is the title of the far-left column on the Analytics page. If looking at a specific time frame (other than "All Time"), frequency trends will be highlighted here.
Level of Interest - A graph that shows the volume of Comments over time (Topics page/Opinions page) and the volume of Records over time (Comments page). On the Topics page and the Opinions page the Level of Interest graph is further broken down by sentiment.
Machine Learning - The study of creating computer programs which perform complex tasks by "learning" a statistical pattern from many thousands of examples, rather than by following preset rules.
Management page - This is where you manage Filtersets and customize Groups.
Mentions - These are contained in a fifth sentiment category. This is where comments that have a Topic but no associated Opinion or single-word responses containing only a Topic reside. This category is useful for tallying the frequency a Topic present in a dataset.
Merge Topics/Opinion - This is used to join two or more Topics, or two or more Opinions, into one group.
Metadata - data that supplements the unstructured text that is analyzed by the software, for example, dates, respondent ID’s and demographics. This metadata can be used for filtering and for Correlations.
Metadata Correlations - Highlights interesting information. These correlations are identified automatically by the system if there's a statistically significant relationship between comments and accompanying metadata. You can also make correlations manually. They are displayed as donut charts and can be exported.
Natural Language Processing (NLP) - The science of designing computer programs that can process or generate human language.
Opinion - A word or words—often an adjective or a verb—that describes an attribute of the Topic it accompanies. For example, in the sentence “The carpet is dirty”, “dirty” is the Opinion.
Organization - A Keatext Organization is like a repository for your datasets. Each company that licences Keatext will be given an Organization. Think of it sort of a company account. A company can have many Organizations. Each Organization has its own Groupings. An Organization can have many users and a user can belong to many Organizations.
Praises - The Sentiment category that contains positive statements.
Problems - The Sentiment category that contains negative statements.
Questions - The Sentiment category that contains questions.
Rank - On the Analytics page, this column ranks results in order of Record or Comment frequency so you don’t get lost as you navigate down the list.
Record - A Record refers to one entire row in the CSV file, regardless of the number of columns/fields/cells present. It is the feedback or response given by a single Commenter.
Record / Comment Count - A column on the Analytics page that displays changes depending on which page has been selected: Comments, Topics or Opinions. It displays the number of records that contain this comment on the Comments page. It displays the number of comments that contain the Topic or Opinion on the corresponding pages.
Record Filters - Filters that are specific to a dataset. They are personalized, based on the data you upload, and they mirror the columns in the CSV file.
Record Filters panel - this opens to reveal a panel allowing you to filter by the metadata contained in your CSV import.
Related Comments - A tab in the Comment Summary that appears when you select a Comment. It contains a few representative comments. To access the original record, click on the text snippet in the Related Comments tab. If you want to see All Related Comments, select the black bar on the bottom of the workspace.
Result (Comment, Topic or Opinion) - This is where the results are displayed on the Analytics page. You can click on a result to view more detailed analytics.
Search Filter (Search bar) - It allows you to search for granular results. Results appear as Keywords, Topics or Opinions.
Sentiment Filters - These filters allow you to choose which categories of statements you’d like to explore: Problems, Praises, Suggestions or Questions.
Sentiment Score - A mathematical calculation that is used to determine the polarity (positive or negative) Sentiment of the Comment. The score will range from 0 to 10.
Settings page - You can manage users and check your usage on this page.
Source Filters - This filter allows you to analyze sources individually, together or in any combination you choose. Activate Source Filters by clicking the checkbox beside each source in the Record Filters panel.
Sources page - This is where you upload files to Keatext.
Suggestions - A Sentiment category that contains recommendations.
Text Analytics - Extracting meaningful insights from unstructured text.
Time Filters - Allows you to isolate and compare time periods.
Tags - Give you greater control over how you organize your data on the platform. You can assemble a combination of Topics, Opinions, and Sentiments together under the same Tag. They can be manually applied to organize your findings further based on KPIs, customer journey steps, milestones, consumer concerns, and so on.
Topic - The main subject of the idea expressed. A Comment may contain several Topics, and a Topic can consist of multiple words.
Total Commenters - The total number of Commenters refers to the total number of people who have given feedback.
UTF-8 encoding - CSV files must be encoded in this standard because it allows for accents and special characters. It is the most widespread encoding, and supports special characters for non-European languages.
Visualizations - The representation of data in graphs.