Social Media Sentiment AnalyticsAnalytics on social media platforms to evaluate and improve brand sentiments using advanced text mining and naturallanguage processing (NLP)

A leading global manufacturer and retailer of beauty and cosmetics products with over $1.15 billion in annual revenue. Currently operates in over 3,000 stores in 61 countries, spanning 25 languages & 12 time zones and sells online in 22 multi-linguistic countries.

Business Objectives

Solutions• V2’s team developed a structured approach to setup connectors to multiple social

media platforms to collect the real-time data.• We extracted essential textual and statistical data from multiple data sources. Then

ingested and merged social, customer and campaign data.• Thereafter we implemented a robust process that standardized and pre- processed

the mined textual data. We then performed various pre-processing and core mining techniques which included:

Benefits Delivered• Our solution helped the client get a sense of how the brand was perceived in the

social world• Sentiment analysis helped gauge customer acceptance of user campaigns• Pre- & Post campaign analytics helped the business stakeholders take informed

decisions for future promotions• Increased sales of brands through targeted promotional strategies• Quicker response times to customer posts, queries, and comments on social

channels

Tools and Platforms Used• Facebook and Twitter Integration via APIs• Customer and Campaign Data integration• Python and related packages Tableau for visualization

• This helped us derive key topics and associated context which were further modeled to arrive at various NLP measures such as relevance, inuence and sentiment polarization.

• The Analytics team assigned a sentiment in three categories – positive, negative, neutral. Data analysis and validation of the model was performed by verifying various trends and patterns. Visual dashboards of key parameters highlighted the pre & post trends and patterns across different segments.

1. Sentence splitting 2. Noun phrase chunking 3. Lemmatization

1. POS tagging2. Tokenization Parsing3. Named entity recognition4. Information Relation extraction

Client wanted to achieve following key objectives:• Measure and manage their online reputation and customer interactions across multiple social media platforms• Measure effectiveness of digital marketing campaigns and their online customer engagement initiatives

And NLP tasks such as: