The Homepage takes the form of a dashboard, with different visualizations of your data. It allows you to quickly and easily spot trends and insights, and helps you narrow down areas that you may want to dive deeper into on the Analysis page. You can create multiple dashboards, based on different metrics, which you can easily toggle between, to gain a fuller picture of your dataset.
The Homepage uses both the metadata contained in your original upload and the analytics applied to the dataset by Keatext, offering you a wealth of visualization possibilities.
You will be taken to the Homepage upon login. To access the Homepage from other pages in the app, select the Homepage icon on the global navigation at the left of the screen.
The Homepage contains configurable widgets allowing you to customize a dashboard that speaks to you! The dashboard is user specific, meaning each person who has access to an organization can save widget configurations according to their own needs. Your dashboard can differ from your colleagues’ dashboard even though you are all members of the same organization.
In the following example we are looking at a dataset containing online reviews of Uber Eats from multiple cities.
When initially generated, the homepage contains 8 widgets. These widgets can be configured, named, resized, and moved according to your needs. Additional widgets can also be added.
The Sentiment filter, located at the top of the screen, allows you to choose which sentiment categories you’d like to explore: Problems, Praises, Suggestions, or Questions. You can mix and match as you see fit!
At the top of the screen you will see the number of Sources you currently have enabled. To see all available sources, click “view sources” just below the number. This will take you back to the Sources page in case you’d like to upload another.
The total number of Records analyzed is indicated right beside the Sources.
The sentiment score is a calculation that compares positive to negative comments to determine overall sentiment.
The Records overview charts the volume of records and sentiment score over time - you can display it by day, week, month, or year.
Volume and Positive vs. Negative ratio bar charts
These two configurable bar charts reveal insights instantly. In the example above, you can see which age range has the highest share of voice - speaking most in online reviews. Move to the Positive vs. Negative ratio chart to see HOW different age ranges are speaking of Uber Eats! If a certain age range has a high ratio of negative comments, you know right away that you have some issues to address with this population.
You can configure the charts to display the metadata variable of your choice, including Tags if you already applied them. Scroll over a bar in the chart to see the total in both frequency and percentage.
The Topics chart allows you to see which Topics are discussed most often, broken down by sentiment, instantly revealing problem areas.
It's clear that there is a lot of negative conversation around the Topic of “fee”. It’s time to tackle that problem!
Top comments heatmap
The Top Comments heatmap shows the intersection between two variables. By default, it is set to display the number of Comments that contain both a given Topic and a given Opinion, but this can be modified using the filters at the top right of the chart. Choose the variables you wish to create an intersection with from the dropdown.
The heatmap can be expanded at the top right of the widget to view more results. The darker the colour, the higher the volume. We can spot the negative conversation around shocking customer service in this example.
To click through to the associated Records, simply click on the number and the Records page appears below.
From there you can drill down to the individual Record to see the original text and all its associated metadata.
If a Record was translated to English from another language you will see a toggle on the top right of the Record Detail. Flip this toggle to see the Record in the original language.
Configuring a widget
Click on the three dots at the top right of the widget to configure, rename, or delete a widget. The configuration popup has two panels. The left panel allows you to set options and filters and the right panel displays any changes made before you update.
It is important to remember each widget is configured separately!
The left panel contains two tabs. The Options tab allows you to name the widget and choose which variable you want to chart. You can choose from the metadata contained in your original upload and the analytics applied to the dataset by Keatext, including Tags, Topics, and Opinions.
The name of the dataset will be written directly below the variable in smaller font. This means you can graph from more than one dataset on the same dashboard! In our example we only have one dataset, Reviews, :/ but you get the idea.
The panel on the right shows you how your widget will be displayed and allows you to limit the number of bars or change the way it is sorted, alphabetically or by volume.
If you don't want to apply any filters select "update widget" at the bottom right and you're ready to go!
The Filters tab allows you to set and save filtersets.
You can choose the Source you'd like to display in the widget, filter by the metadata contained in your dataset using Record filters, as well as filter by Tags (if you have previously applied them), and even use the Search filter to isolate Topics, Opinions, or Keywords.
If your dataset includes dates, choose that variable in your Records filter to isolate time periods.
Once you have set all your desired filters, select "Update Widget" on the bottom right of the popup. This means your widget has been configured with the chosen parameters but you have not saved the filterset to use it later or in other widgets.
If you feel you will need to reuse the parameters multiple times select “save new filterset” to come back to it easily at any time. Toggle between saved filtersets with ease using the dropdown. You can even pick your saved filtersets in other widgets. if there is a filterset that you will use often it is probably better to save it rather than simply updating the widget without saving.
If you modify a filterset you have already saved, and you want to update it so it is saved with the new changes, select “update everywhere”.
Again, it is important to remember each widget is configured separately.
Renaming a widget
To rename a widget without reconfiguring it, select the three dots on the top left of the widget, type the desired name in the field and then hit the “Rename” button directly to the right of the field.
Adding a widget
You can add additional widgets by selecting the “Add Widget” button at the top right and then choosing a style of widget from the pop up. Configure your widget in the same manner as detailed above.
Editing the layout
You may want your widgets arranged in a different order on the default dashboard. Select “Edit Layout” at the top right of the screen to do this.
Place your mouse over the two lines beside the name of the widget and drag and drop it where you see fit.
To change the size of a widget use the three dots on the edges.
Duplicating, Deleting, and Sharing a dashboard
Select the three dots on the top left of your dashboard to duplicate, delete or share your dashboard.
Sharing a dashboard is useful so that stakeholders who don't have access to the Keatext app can see the dashboards and even click through to the records. It acts as a sort of "report".
Duplicating a dashboard is useful is you want to make multiple dashboards with the same layout. Simply duplicate the dashboard and change the configurations accordingly.
You can always delete a dashboard of it is no longer necessary.
To add a new dashboard simply scroll over the homepage icon and select the "+ new dashboard" button. You will be taken to a completely empty dashboard. Here you can add your widgets and design your dashboard from scratch. To toggle between different dashboards simply select the desired dashboard from the popup that appears to the right of the icon.
Creating multiple dashboards based on different segmentation, like we see below, can help you easily see your data visualized from different perspectives. Here we have one dashboard that exclusively displays the insights related to the customer journey step - Delivery. We can also look at the data segmented by city, in this case Toronto.
The Homepage offers you insights at a glance. If you want to dive deeper into any of the issues you see here, move to the Analysis page and continue text mining!