Using Filters to refine your dataset

1. What are 'Filters' in Wordnerds?
2. Selecting and applying filters
3. Saving and editing filters 
4. Using the Keyword Search filter
5. Using the Date Range filter 
6. Using the Sentiment filter 
7. Creating a new filter 

---

 

What are 'Filters' in Wordnerds? 

Filters are basically the traits tied to the content data you want to check out in your project. When you use them, they help you fine-tune your dataset in the best way possible.

This means you can benefit from having all of your data in one place, but with the flexibility to narrow down to a specific dataset when needed. And, because you are able to create custom 'Themes', you'll find that you're able to use these to really customise how your data is filtered. 

In the Wordnerds platform, you can refine your dataset by...

Universal Filters:

• Keyword Search (without needing to create a Keyword Theme)
• Date Range 
• 'Region', 'Brand' and 'Competitors' (these are set at the data input/upload stage)
• Sentiment

Custom Filters:

• Upload Themes (metadata you have included with your data input/upload) 
• Context and Keyword Themes (classifications you have created)
  
  

---

 

How to select and apply filters

You will be prompted to apply filters when you enter any analysis feature in the platform. By default, the filters will be set to include all data from the previous month. 

The filter page

As well as the universal filters, the filter page will show the Theme Categories, and Themes within them, that make up the custom filters that are specific to your project. At the top, in the filter bar, you can see the filters that have been applied based on your selections. 

The filter page will look something like this: 

This article will walk through using each of these filter options. 

Filtering by Keyword Search

Where you have the option to filter by Keyword Search, this can be great for getting results quickly that contain specific words or phrases, without needing to create a Keyword Theme. 

👉 Read more about filtering by Keyword Search 

Filtering by Date Range

By default, the date range is set to the previous month, but is easily changed. You can choose from the options on the drop-down, a custom period based on calendar or fiscal periods - or you can filter by custom date range. 

👉 Read more about filtering by Date Range

Filtering by Sentiment 

'Trending', 'Topic Analysis' and 'Inbox' allow you to filter by sentiment. Simply make your selection using the sentiment emoticons. 

👉 Read more about filtering by Sentiment

Filtering by 'Regions', 'Brands', 'Competitors' and 'Themes'

'Regions', 'Brands' and 'Competitors' are determined at the data input/upload stage.

The other filters shown on the filter page are based on your 'Themes'.

For 'Regions', 'Brands', 'Competitors' and 'Themes', you can filter at category level or individually.

Filtering at category-level

To filter at category level, simply use the toggle next to the category name. It will look something like this: 

In this example, I have toggled on the entire 'Star Rating' category, and all filters within this category have been selected as a result. I can also see the number of individual filters that have been selected ('5 selected'). 

To select filters individually from the SAME category...

Expand the category using the down icon and toggle the individual filters.

Selecting multiple filters within the same category will look something like this: 

In this example, I have chosen to select the '1' and '2' filters in the 'Star Rating' category ('2 selected'). Because these filters are within the same category, once they are applied, my results will show data matched to either the '1' OR '2' filter. 

Selecting individual filters from categories with over 10 filters...

If there are 10+ filters within a category, instead of the toggle off/on view, the filters will show as a drop-down, sorted alphabetically. Click on the filters you want to select, and you'll see them move to the filter selected bar. It will look something like this: 

In this example, I have selected three filters from within the 'PURCHASE | Shopping' category ('3 selected'). Because these filters are within the same category, once they are applied, my results will show data matched to the Product Availability' OR 'Product Variety' OR 'Product Quality' filter. 

To select multiple filters from DIFFERENT categories...

Follow the same steps as above to select either whole categories or individual filters across the different categories. 

Selecting multiple filters from different categories will look something like this: 

In this example, once applied, my results will show data that is matched to any of the 3 filters selected in the PURCHASE | Shopping category ('Product Availability', OR 'Product Variety', OR 'Product Quality') AND that is also matched to any of the filters selected in the Star Rating category ('1' OR '2').

Confirming and applying your filter selection...

After completing your filter selection, scroll down to the bar at the bottom of the filter page. Next to the FILTER button, you will see an indication on how long it will take for your filters to be applied. It will look something like this on your screen: 

Applying your filters won't usually take long (a few seconds) BUT, if it is a particularly large data set (we're talking hundreds of thousands, or millions, of data points), you'll get a heads up at this stage that it will take some time to process, which gives you the opportunity to refine the dataset further by applying more filters before confirming and waiting. 

Filtering to Include or Exclude Spam

Content from 'Authors' (unique users) who have 15 or more pieces of verbatim attributed to them in a 24-hour period are marked as 'Spam Authors' in the Wordnerds platform, and you can opt to ignore their content by keeping the 'Include Spam' filter toggled off.

Filtering to Include or Exclude Automated Clusters

Automated Clusters are groupings of 'Topics' that are deemed to have similar or overlapping meaning. If you filter to include Automated Clusters you'll see 'Topics' have been grouped in your results.

👉 Read more about Automated Clusters

When you are ready to confirm and apply your filters, click the FILTER button.

 

---

 

 

Saving and editing filters 

Once you have applied your filters, they will be saved for your session. This includes if you switch between analysis tools, saving you from the need to reapply the same filters each time. 

You'll always be able to see what filters are applied by viewing the filter bar at the top of your screen. The filter bar will look something like this: 

To edit the filters that have been applied, click the Edit Filters button on the filter bar, or click the 'X' icon next to the specific filter if you want to remove it; either action will take you back to the filter page to make and apply your changes. 

---

 

How to use the Keyword Search filter

Keyword Search is great for...

• Quickly and easily filtering to specific words or phrases as a one-off, without needing to build a keyword theme
• Exploring how frequently specific words or phrases are used in your data to decide if it is worthwhile building a keyword theme to classify and track an issue based on that vocabulary 
• Exploring the different ways specific words or phrases are used in your data to help you train a new context theme to classify and track an issue based on the context in which that vocabulary is being used 

To create a Keyword Search filter...

• Create your filter rules by typing in the words and phrases you want to filter by in the Includes bar, pressing ENTER after each word or phrase to separate them
• From the drop-down on the Includes bar, select whether you want your rules to work on an 'AND' or 'OR' basis
• If you are aware of ways that the keywords or phrases you have put in the Includes bar could be used in a way that you don't want to include, add to your filter rules by typing in the words and phrases you want to exclude in the Excludes bar, again pressing ENTER after each word or phrase to separate them

The Keyword Search filter will look something like this: 

In this example, I have chosen to filter to posts that contain the keywords 'cat' OR 'dog, but do not also contain the word 'food'. 

---

 

 

Using the Date Range filter

By default, the date range filter will be set to 'Previous Month'. To edit this, expand the filter using the down arrow. You can then choose to select a date range from the drop-down options, or to select a custom date range. 

The drop-down date range options will look like this: 

To filter by a custom date range, choose Custom from the drop-down options.

From here, you can select custom calendar periods, or, if you have added your organisation's fiscal calendar to the project settings, by custom fiscal periods. 

👉 Upload your fiscal calendar to your project settings

The custom calendar periods options will look something like this: 

To filter by a custom date range, click the Custom dates button/calendar icon and enter your date range using the calendar. Your screen will look something like this: 

Click Done once you have selected the Date Range filter you want to apply.

Comparing periods

In 'Trending' and 'Deep Dive', you will get an automatic comparison with the previous equivalent period, customised to the date range filter you apply. For example, if you select 'Previous quarter', your data will automatically be compared (in terms of volume and sentiment) with the quarter before that, but if you select 'Previous week', you data will automatically be compared with the week before that, and so on. 

The date range comparison will look something like this (view from 'Trending'): 

---

 

Using the Sentiment Filter 

'Trending', 'Topic Analysis' and 'Inbox' allow you to filter by sentiment.

Filtering by sentiment is great for...

• Focusing your analysis on a particular sentiment group, for example, content that is very negative in sentiment. 

To filter by sentiment, simply toggle the sentiment category/categories you want to filter to using the emotion/s (ranging from very positive to very negative).

The sentiment filter will look something like this: 

👉 What is sentiment analysis? 

---

 

Automated Clusters

Automated Clusters group together 'Topics' that appear to have similar meaning. When selecting your filters, you can choose where to show your results with 'Topics' clusters together, or not.

You can also toggle automated clusters on/off in your view, using the toggle icon. 

If you choose to select Automated Clusters, they will look something like this (view from 'Trending'): 

In this example, you can see that automated clusters have been toggled 'on', as indicated by the toggle icon next to the topic groupings. You can expand a topic cluster to see the topics within, as shown here ('chat-live', 'conversation-live', and 'live-talk' have been grouped under the topic cluster 'live chat'). 

Creating a new filter 

If you want to filter your data using a filter that does not currently exist, you will need to add a new 'Theme'.

👉 What are 'Themes' in Wordnerds

If the new filter is based on an Upload Theme (i.e. metadata)... 

For example, if you want to filter by demographic or segmentation data, or a NPS score or group, this can be added as an Upload Theme during your data input/upload and then used as a filter. 

If you data is feeding in via API, you will need make these changes at source. 

👉 Connecting your data to Wordnerds via API

If your data is being manually uploaded by CSV, you can simply add in the new metadata column on your next upload to have the filter going forward. As soon as the data upload is processed, the new Upload Theme will be available on the filter page. 

👉 Add your data to Wordnerds via .CSV Upload

To retrospectively apply your new Upload Theme (which can then be used as a filter) to data that already exists in your project, you will need to reupload the data with the new metadata attached, and then delete the original upload to avoid duplicating the data. 

If the new filter you want is based on a Context or Keyword Theme (i.e. is a way of classifying your content data)...

This is much easier! Simply create your new custom classification by creating a Context or a Keyword theme. When you enable/save your new theme, you can opt to apply it historically, and, once your new theme has processed, it will be available to use on the filter page. 

👉 Add a new theme

👉 Train a context theme

---

Anyquestions? 

We’re always happy to chat. Reach out using the help bubble at the bottom of your screen or contact your Customer Success Manager.

Article written by: Lyndsay, Senior Customer Success Manager