- Customer is one of the big gas producer. Had failure in well pumping system frequently due to different issues
- BrainWave Consulting was asked to provide Realtime failure warning & prediction system
Case Studies: Data Analytics
Oil & Gas field Pump Failure Prediction
Customer Situation
Solution Steps
- BrainWave Consulting partnered with domain experts to understand the causes
- Built algorithms to clean and annotate data for learning
- Built custom Machine Learning algorithms to predict data Realtime
- Built alert system to provide alerts to user on an incidence
Benefits
- Reduced operations effort in identifying failures
- Early warning to help take corrective actions

Predictive Analytics for Leakage for O&G Major
Business Challenge
- Sub-sea leakage is very difficult & costly to identify by physical verification
- Heavy production loss due to leakage point and delay in detection
- From sensors multiple data points available but no specific pointer
- No clear pattern to identify leakage based on linear curve
Solution
- Work with business user to understand scenario
- Define relevant model to be used for required analysis
- Study and run test on previous occurrence of leakages
- Use of R and Shiny along with Greenplum DB
Benefits
- Flexibility to generate on demand reporting enabling user to predict and take preventive steps
- Provide graphs and charts for easy interpretation by business users
- Generate patterns and exceptions based on statistical models
- Cost saving in terms of prevention and early detection of leakage

Document Processing Using ML/AI for Airline
Business Challenge
- Detection of fraudulent Invoices
- Extraction of Invoice data from different file formats & different European Languages
- Each vendors has unique format, quick onboarding of vendor
- Payment delay attributing to invoice processing affects operations
Solution
- Intelligent algorithms, filters used to find specific data fields in variable locations
- Annotating each extract line-items from invoices with detailed data about the merchandise included in an invoice (including tables)
Benefits
- Intelligent identification of Duplicate Invoice Information
- Fraudulent invoice entry detection
- Three way auto validation
- Digital Invoice Processing – Read and Scan invoices
- Capture and classify invoice data elements with intelligent logic
- Ability to avoiding late payment fees and potential for capturing early payment discounts from suppliers

Knowledge Discovery from Unstructured File system
Business Challenge
- Large volume of files gathered over a period from multiple users
- Un-organized, unstructured file system
- Need faster analysis of content and categorization
- Generation of extract with new Taxonomy & Classification after due clean up
Solution
- Provide search recursively within content / folders/ file names
- Assign relevance in terms of score for sorting
- Facility to narrow result as required
- Web based system – Search by multiple user
- Export data in XLS in desired format
Benefits
- Ability to analyze and clean up files system
- Provide taxonomy and classification as to avoid further manual operation
- Address all files types to avoid manual work for exceptions
- User defined extract to be effectively used further for knowledge management

IOT based Water Quality Monitoring
Business Challenge
- Affluent generated by Pharma manufacturing process is highly contaminated, such water can’t be release in soil or water body
- Any contamination in water will lead to penalty by pollution board
- Affluent is generated 24X7 in plant operation
- Needs early warning if water treatment plant is not working
Solution Approach
- Read sensors from water treatment plant every few minutes
- Alert sent to operation head if sensors are not working properly
- Model based prediction for plant failures
Benefits
- Improved Plant operational availability
- Effective plant maintenance management
- Improved regulatory compliance

Sentiment Analysis of Online Customer Review for Product Developers and Marketers
Business Challenge
- Understanding of Customers for the Product / Service
- Measure relevance of Product / Service for the customer
- Analyze customer feedback for specific aspects of Product / Service
- Unstructured feedback makes it difficult to analyze
Solution
- ML based Sentiment and Aspect analysis of the customer feedback
- Solution for verticals such as Insurance, Healthcare, Personal care and others
- Ability to incremental learning for the solution based on review and edits
- Provides actionable dashboards, ability to search and review for further analysis
- Custom solution based on best suitable model for the review analysis
Benefits
- ML based Sentiment and Aspect analysis of the customer feedback
- To identify strengths and weaknesses of the specific product / service features
- To address customer grievances in specific area of concern
- To come up and track effective marketing campaigns, for effective marketing ROI
- Better customer engagement resulting in increased market share
