Business Use Cases

Business use cases with analyses are illustrated here.

Five Business Use Cases

Case 1: Tweet Sentimental Analysis Using TextBlob And NLP

NLP powered tweet analysis

  • Sentiment analysis is the task of determining the emotional value of a given expression in natural language.
  • Using collected tweets data per brand, perform sentimental analysis using - TextBlob
  • Textblob (NLP Api) sentiment analyzer returns two properties for a given input sentence:
    • Polarity: -1 is negative, +1 is positive sentiment
    • Subjectivity: Subjective sentences generally refer to personal opinion, emotion, or judgment. Use polarity to assign word to each tweet: positive, negative and neutral

Clients can get insights from sentimental analysis of customers’ tweets.

  • This graph shows the relative perception of brands based on tweet sentiment analysis
  • Again, tweets are processed through an NLP model and classified if they were either: Positive, Neutral or Negative in sentiment
  • For example, GLAMGLOW has received a lot of flak and bad press KORRES on the other hand has generated a lot of positive buzz on Twitter

Case 2.1: Key Performance Brand Level

  • This graph shows the relative perception of brands based on tweet sentiment analysis
  • Again, tweets are processed through an NLP model and classified if they were either: Positive, Neutral or Negative in sentiment
  • For example, GLAMGLOW has received a lot of flak and bad press KORRES on the other hand has generated a lot of positive buzz on Twitter

Case 2.2: Key Performance Product Level

  • This graph shows the relative perception of brands based on tweet sentiment analysis
  • Again, tweets are processed through an NLP model and classified if they were either: Positive, Neutral or Negative in sentiment
  • For example, GLAMGLOW has received a lot of flak and bad press KORRES on the other hand has generated a lot of positive buzz on Twitter

Case 3: User Journey On Demo App Built On Tkinter (Python GUI)

  1. User opens the application
  2. User searches for the interested product
  3. Search query result is displayed with brand logo and product photo

Shows the product’s ingredients and whether it contains harmful ingredients with a kind warning.​

Users can get comprehensive information on interested products’ ingredients and whether it contains any harmful substance

Business Case Convert users to purchasing products by becoming go-to place for searching cosmetics products Affiliate revenue is sizable with growing user base

Case 4: Competitor Lookup

Clients can get a comprehensive overview of competitor brands.

  • Used Tableau to create a Brand Dashboard
  • Dashboard uses brand name as a filter so that the user can choose which brand to look into
  • Dashboard contains information like brand description and their products
  • Dashboard also shows the brand logo and its official website if present

Case 5.1: Customer Demographic and Attribute Distribution

Clients can understand customers based on user review analysis.

  • The brand’s main audience are between 18 to 34, although there are also customers from other age bands
  • Most customers are lighter skinned
  • Eye color distribution show that they have an area of improvement where they can release products that may match better with gray eyes

Case 5.2: Average Product Rating Over Time

Some products might be seasonal.

  • Product review started off strong, but tapers off with slower frequency of review and dropping average rating value
  • Might suggest that product is not as good as expected or longer-term issues with the product that wasn’t immediately apparent (e.g., weather incompatibility)

Case 5.3: Product for Specific Area of Concern

Review data can tell clients where users find the product most useful.

  • “Purity Made Simple Cleanser” was made for acne problems in mind
  • May also help with aging skin
  • Not a good fit for customers who are looking to fix stretch marks or sun damage
  • Brands can know consumers’ pain points and use insights to target marketing and improve formula to strengthen the product positioning
  • Consumer can make better informed decision about their intended purchasing before paying

Case 5.4: Top Products with Certain Attributes In Mind

Review data can tell clients all time favourites based on users’ attributes.

  • For undecided customers, they can first look up for items that are often used by others who share similar attributes
  • For example, they may have a “combination” skin type, light skin tone, and blue eye color
  • This insight can help them choose the right product in a specific category (e.g., foundation)
  • Graph shows each product’s “love” rating (a measure of how many times customers “love” a product), and their numeric rating
← Result
Models →