Spectra Analytics - Consulting Services

Overview

Spectra Analytics works across a broad range of industries from healthcare to marketing and finance. We aim to empower businesses by generating actionable insights from their data.

Our bespoke solutions may be implemented in new or existing proprietary systems or increase the flexibility and sophistication of standardised platforms.

Data Management

This includes database management, data cleaning and enrichment from public and private sources (where possible) including social media

Data Analysis

Providing business insights through mathematical and statistical modelling and machine learning.

Research

Develop cutting-edge research for academia and business

Advisory & Training

Provide consultancy services and training for businesses on data analytics

Spectra work on fixed term projects and on a consultancy basis. Please contact us to see how we can help.

Case Study - Analytics Dashboards

Hospitality

The hospitality industry produces large amounts of data on people's travel and interests. This is particulalry, true of Online Travel Agents such as Expedia and Booking.com who track hotel and flight prices in real time. This data is helping consumers find better accomdation at better prices but it is also helping hotels connect with their target customers and monitor their competition.

We worked closely with Bloc, a small upscale travel hotel chain, to try to make sense of all their data by developing an Analytics Dashboard. The system monitors and analyses business metrics in real time and improves efficiency and productivity by automating report generation.

The system incorporates machine learning to analyse data from a wide variety of sources - such as the property management system, online travel agents, airports, local events, online reviews, Google Analytics, Google Adwords and weather - to forecast demand; predictive analytics. It then uses this information to optimise room pricing to maximise revenues; presecriptive analytics.

Bloc

Case Study - Robotic Process Automation

Digital Transformation

The 4th Indutrial Revolution is ushering in a new era of automation which is impacting every aspect of business. This is a huge challenge for companies as they look to not only change their digital infrastruture but also change their company culture. Companies must become data led not just using data to support their decisions.

We worked with S6 - the UK’s largest network of FA-affiliated 5 & 6-a-side football leagues with over 30k members - to implement a large scale digital transformation programme.

Spectra developed a web app to automate leagues and team management and improve finance reconcilliation. The system also generates real time KPI’s to monitor live business performance. We also integrated marketing and customer interactions seamlessly with social media.

S6

Case Study - Text / Document Analysis

Healthcare

Healthcare is one of the major producers (and consumers) of data. Data is constantly being produced from real-time medical recording equipment and health monitoring services. Medical facilities have huge databases of patient records. The amount of medical data is likely to explode over the next few years with wearable technology incorporating health monitors and the expansion of genome and exome sequencing for personalised medicine. This data is being used to by medical facilities to monitor and treat patients and by governmental departments for oversight.

Spectra are using Natural Language Processing to analyse text in medical documents and GP notes from the Electronic Health Record. This research is being used to improve patient triage processes in primary care with the aim of reducing GP workload. This project is being conducted in conjuction with the University of Manchester and funded by InnovateUK.

The algorithm uses word embeddings, such as Word2Vec, to convert text in to vector representations. These are then combined with a deep learning classification algorithm to predict READ and SNOMED codes which represent patient symptoms and treatments.

Case Study - Image / Video Analysis

Transport

It is predicted that the Automotive industry will be the 2nd largest generator of data by 2015. They produce an abundance of data from vehicle and manufacturing sensors to customer and dealer information. They are using this data and analytics to produce better vehicles and develop better insights into their customers and businesses. Transport authorities are also utilising data to better manage and deploy public services.

Spectra have developed algorithms to analyse traffic camera footage. The algorithm analyses the images in real-time to detect the object type and its location. Through this analysis it can detect traffic flows and near miss / actual collisions. It is based upon deep learning methodologies enhanced by inference techniques. This is a far more accurate and lower cost alternative to human enumerators which currently conduct this analysis.

This project is currently being tested as part of a POC to evaluate GPU compute requirements with our partner UKCloud

Deep-i

Case Study - Regulatory Reporting

Compliance

Regulatory changes are making significant changes in the way in which financial institutions function. It is hoped that this will lead to greater transparency, improved consumer choice and increased stability. Unfortunately, these requirements have created a huge burden on financial institutions as they attempt to modify their business models to cope with the new regulatory environment.

Spectra worked closely with the financial brokerage firm XBZ to automate their processes to comply with MiFID II regulatory reporting requirements. This entailed integrating and normalising trade data from multiple sources and generating tailored reports for regulatory authorities.

Case Study - Surveillance

Finance

Over recent years there has been a number of high profile cases of illegal trading and market manipulation. In the UK, Navinder Singh Sarao dubbed the 'Hound of Hounslow', used a method known as 'spoofing' to illegally generate tens of millions of dollars. It has so far been very difficult to identify and prosecute offenders due to the sheer volume of trades and the complexity of the criminal activities.

Spectra developed machine learning algorithms to detect illegal trading and market manipulation in the financial markets.

Types of illegal activity include:

  • Spoofing
  • Layering
  • Insider trading
  • Front Running
TAD

Case Study - Survey Analysis

Healthcare

Spectra worked with the Liverpool School of Tropical Medicine on behalf of Unicef to develop a statistical tool for public health teams to design Lot Quality Assurance Sampling surveys. LQAS is a survey design which uses stratified sampling to label regions as having either ‘acceptable’ or ‘unacceptable’ levels of an indicator. They are cheaper to perform than a traditional cluster surveys and regions can be combined to give point estimates. This work was then extend to Multiple Category LQAS and Large Country LQAS. These tools are available below:

LSTM

Case Study - Spatial Analysis

Accessibility

Transport authorites spend a lot of time trying to improve accessibility but without good quality data this can be a big challenge. Spectra developed new analytics tools to study accessibility on the London transport network for mobility-impaired users. The aim was to identify 'mobility black spots' by calculating the time differentials between TfL recommended journey times and those of mobility impaired users. This information could then be raised to the relevant transport authorities. This was done by analysing the GPS data of mobile phones of mobility-impaired volunteers.

The project was funded by a European Commission Grant in association with Organicity

LSTM