Aspect_Sentiment_Analysis

Use Case: Customer Service Sentiment Analysis

Using review data and feedback from consumers online, we analyse over 32,000 reviews towards the top 5 internet service providers (ISP) within the UK market.

This analysis is carried out following our three-step augmented analytics approach: Gather, Process and Visualise.

Read More
Consumer behaviour for online shoppers

Understanding Consumer Behaviour: Causal Inference and Sentiment Analysis

Data has a vital role in marketing, but there is much more to discover when it comes to data-driven approaches. 2020 Insights and Data Analytics Trends revealed that new investments will be made into AI for smart insights due to the fact that AI helps break through confusing, and at times conflicting, observational data. AI…

Read More
MDUK document tagging project

Muscular Dystrophy UK

Saving researchers valuable time Auto-tagging and classifying research papers and articles that lets patients, researchers and staff quickly find relevant information. The challenge Suzannah, Head of Digital at Muscular Dystrophy UK, and her team needed help with the organisation of their website content in preparation for a makeover. The project included 7,961 pages from their…

Read More

Deloitte

Turning news into sales opportunities Account managers at Deloitte close more business thanks to actionable insights delivered straight to their phones.  Keeping them current with customers and ahead of the competition. The challenge Deloitte’s partners and account managers found they were drowning in news from sales support teams, and unable to react quickly to market…

Read More
Whitespace with deeper insights

Whitespace discovery for consumer goods

Market Research: Whitespace Discovery We take a look at how product managers for the niche Robo Lawn Mower category, can analyse consumer feedback to discover whitespace in an emerging product category. With a rapidly evolving category with new entrants and evolving consumer requirements, it’s hard to know what insights to look for and where the…

Read More
Location ID machine learning model

Location

What does this model do? Identify geographical locations – countries, cities, oceans, rivers, etc. – found in a body of text or a document. The entities are normalised to Latitude, Longitude and Elevation – and include the Geonames Database location ID, population estimates etc. Who should use this model? Analysts and Researchers that are interested…

Read More
company distress machine learning model

Company Distress Prediction

What does this model do? A prediction model that identifies stress indicators in financial reports, CEO letters and earnings transcripts. Trained on a large corpus of financial documents from the NASDAQ, NYSE and FTSE and backtested over the last three years. Works in English language only. Who should use this model? This model works very…

Read More

Semantic Role Labelling

What does this model do? This model is used to identify the role one or more objects has with the other(s). Take for example a company team page that shows members, and you want to identify the job roles of certain people. This model identifies the individual person and the job they have. Such as…

Read More
Topic discovery machine learning model

Topic Discovery

What does this model do? Used to discover the unknown topics within big data. Unsupervised learning technique that converts a large corpus into a network representation, and then identifies and analyses communities (aka topics) in this network.   Who should use this model? This is the ideal discovery tool for those that don’t know what…

Read More
feature recognition machine learning model

Feature Recognition

What does this model do? This model recognizes features of a product, for example a car will have features such as size of its tank, number of seats, media system, alloy wheels. These are all recognized features by the model. Or The model will recognize aspects of a service. Such as an automotive insurance company…

Read More
entity recognition

Named Entity Recognition

What does this model do? This model recognizes; brand, company, product and peoples names within text. Who should use this model? This model has many applications and can be combined with a range of Market Research and Business Analytics use cases. Ideally combined with models such as Sentiment, Event detection or Company Distress. Business Analysts:…

Read More
Consumer feedback model

Consumer Feedback

What does this model do? This model uses review data, comments, blogs and forum data to extract the explanation for the sentiment expressed towards a given aspect of a service or product feature. Combined with our Consumer Sentiment model, you can use this model to identify the reasons why a consumer has given such negative…

Read More
Consumer sentiment model

Consumer Sentiment

What does this model do? Identifies the polarity or sentiment associated with each aspect or feature mentioned in the input review text. Who should use this model? Used for Market Research in analysis of consumer insights, customer experience and consumer behaviour. The Target Sentiment model identifies the specific features of a product or aspects of…

Read More
Event Detection model

Event Detection

What does this model do? Classifies the topic of the news story or article provided in the target webpage into one or more of the following topics: Acquisitions Partnerships Personnel Change Investments (More can be added on request) Who should use this model? These event detection models can be used for media monitoring, sales or…

Read More
News summarisation

News Summarisation

News Summarisation What does this model do? Performs extractive summarisation of the input text, i.e., it identifies the most important sentences from the source text, and generates a summary by sequentially concatenating them. Who should use this model? This model performs automatic summarisation of news articles, blogs or other types of written content, such as…

Read More