TEXTIENT DATASET is an important part in deriving qualitative Cognitive Analytic Insights viz, TEXTIENT BRAND-ESSENCE and TEXTIENT CONSUMER DNA. These insights are generated by running Cognitive analytics on the TEXTIENT Platform using DATASET.
TEXTIENT BRAND-ESSENCE and TEXTIENT CONSUMER DNA provides strategic Brand, marketing and Consumer DNA insights from social media communications/content mined from Facebook, Twitter or YouTube. The insights are predicted by deciphering consumer psychology and behavioural aspects using IBM Watson’s cognitive technologies.
Once you log-in to the TEXTIENT using your Email Address and password, you will be taken to the Dashboard. In the Dashboard you can perform the following tasks.
1. Create A Dataset.
- Create a dataset for a brand, product or service from the social media touch points viz Facebook, Twitter, Youtube or from your own data.
- View the datasets those you have created.
- Get details about specific dataset.
- Perform Actions on a specific dataset.
2. Generate Insights.
- Create the TEXTIENT Brand-Essence, Negative Brand Impact, Consumer DNA and Negative Consumer DNA Insights report using the Dataset you have created.
- View and perform actions on Insights reports.
3. View Your TEXTIENT Insights.
- View your Insights reports.
- Get details about specific Insights report.
- Perform Actions on a specific Insights report.
Please find below the screenshot of Dashboard.
From the Dashboard, you can get started using the following steps to create a dataset.
- Step-1: Here we define the DATASET parameters including dataset sources to create the dataset. Next, We execute the Data mining, data processing and the DATASET gets created. (Run the data acquisition and processing job to create a dataset). Use this DATASET to generate the TEXTIENT BRAND-ESSENCE, NEGATIVE BRAND IMPACT, CONSUMER DNA and NEGATIVE CONSUMER DNA Insights reports.
- Step-2: Here we can access the created DATASET and perform Actions like Get Details of the Dataset, View/Filter Dataset, View Sentiment Analysis and Delete Dataset.
Here we define the DATASET parameters including dataset sources to create the dataset. The Datamining and data processing are run to create the DATASET that will be used to create the TEXTIENT BRAND-ESSENCE, NEGATIVE BRAND IMPACT, CONSUMER DNA and NEGATIVE CONSUMER DNA Insights reports.
E.g. We want to do Brand Research / Analytics from Audi USA’s Facebook page .
From Dashboard, click on the ‘DATASETS’ tab on the left hand side of the page. You will be taken to the “Datasets” Page. Click on the Green coloured “Create Dataset” button. You will be taken to the “Create Dataset” page. Enter the following details.
- Provide a name to the DATASET (e.g. AUDI A4 DSI )
- Provide tags to identify or associate the dataset (e.g. auto, audi, facebook, test)
- In the “Datasource” dropdown, click on any of the following elements – Bring-Your-Own-Data, Facebook, Twitter, Youtube and select the Datasource for creating DATASET. Please refer the following screenshot.
- Enter specific Facebook page(s) or Youtube Video URL(s) or if you use twitter provide query parameters after checking twitter (e.g. hashtag: #audiusa and search term: “audi car” and twitter handle @audi).
- Select a facebook URL or Youtube URL which are fairly popular. Pages or videos which are not popular in terms of likes or viewer counts may not have enough comments for analysis.
- In the Youtube Video , at the bottom, you can see the no. of comments . select videos having atleast 250 comments.
- In facebook, select popular pages where we can get a sense of people commenting.
- In case you are using Bring-Your-Own-Data, click on the “Select & Upload Dataset CSV File” in the “Data source” dropdown.
- In this example, lets use this AUDI USA facebook page : https://www.facebook.com/audi/?fref=nf
- Note: CSV files must be 2-column – (name,message) – formatted with UTF-8 encoding. For help on how to create files with the correct format please refer the following link (Learn More )
- Select the date range for which you want to create the DATASET (From This Date, To This Date). The minimum date range supported is 8 days.Enter “Mask” terms (optional) in the Text field – “Mask These Terms If It Occurs In The Message (Mask Words)”. This masks the messages (consumer comments) containing the words you specify. For More info on selection of Mask Words, please click on More Info.
- Enter “Reject” terms (optional) in the Text field – “Reject Messages Containing Any of These Words (Reject Words)”. This rejects the messages (consumer comments) containing the words you specify. For More info on selection of Mask Words, please click on More Info.
- For example: wallpaper picture pictures video deals why how etc. For instance, why do you want to drive down to New York or say a comment I took 20 pictures of this car with my daughter is not relevant to our analysis which can be rejected. So quickly scan the Facebook or YouTube pages and you may decide on the suitable reject words to use.
- Once you enter these information, please REVIEW all your entered/selected infromation from top of the current page. If the displayed information in the page is correct and as per you need, click on the Green coloured “Create ” button. This will initiate a Job to fetch the data, mine, analyse, spam-filter and finally creates a dataset that represents the minds of hearts of the consumers (human-truth) .
Please refer to the screenshot below which shows sample DATASET creation steps.
Once the Dataset creation job is started, the TEXTIENT platform acquires data and processes it – including Spam-Filter, Data mining etc and then creates a DATASET.
Please refer the below screenshot which shows the Dataset creation job status.
The DATASET will be created in 30 to 45 mins’ time. Once it gets created successfully, you will get a notification on your e-mail.
If the DATASET creation job is not successful, you will receive an e-mail mentioning that the Job is not successful for Technical reasons or Insufficient Data within the given date range.
Once the DATASET creation job that you started gets successfully completed, you should be able to see the dataset you had created under the DATASET history.
Please refer to the screenshot below.
- In the DATASETS page, locate your DATASET.
- Within your displayed DATASET, click on the “Details” button to view its details.
- Click on the “Actions” button within your displayed DATASET, to perform any of the following ACTIONS.
Filter Data by selecting “Mask” or “Reject” words for the 3 tabs – “Low Distribution Two Word Occurrence”, “Top Two Word Occurrence”, “Top Single Word Occurrence”; Review them via the tab “Review” ; Click on the Green coloured “Cleanup” button to perform the DATASET Cleanup. Please refer the screenshots below..
View Sentiment Analysis of the DATASET.
You may decide to delete the DATASET if it was incorrectly created or for other reasons. Click on “Delete Dataset” for the same.
Please refer the screenshots below.
Please refer the video below to understand the above steps.