Our website uses cookies to improve your online experience. By continuing to use our website you consent to cookies being used according to our Privacy Policy.


Hiring a Freelance Data Scientist

Why Data Scientists are a must to improve any TMT Business

More than most industries today, the TMT sector is a naturally digital domain. In the telecom space, voice calls are transmitted digitally, data services often outweigh voice services in importance and revenue, and each year more of the industry’s infrastructure is digital and software-driven. But the boundaries across telecom are diluting into media and ICT acting as digital transformation enablers and drivers. This includes all areas of the business from customer facing areas such as sales, service delivery and customer service as well as back-office operations.

As data scientists, our job is to extract business insights from noisy data

Data Science and AI are vital to the sector and becoming more so as the industry evolves. In the specific case of telecom services, the network operations, customer service operations, and infrastructure operations all generate massive amounts of data. This is why data science in telecom is now so prevalent. Data science and AI in telecom provide operators with the tools to interpret that data and use it to increase reliability, decrease costs, and improve customer service — and much more.

Find and hire the best and most qualified data scientist freelancers through Outvise paying only for the time/project you need them.

Certified Data scientists in the Network

case study
Data science case study in a Telecom company

Challenge, Context, Problems to be solved

Implementation of machine learning models in order to detect anomalies in the customer behaviour in the Telecom context.

Mission, tools and methodology

Telecom operators often have customers that use their subscriptions disproportionately leading to diminished network experience for others. Some examples are customers on unlimited plans with steep data downloads, leading to network congestion or customers who primarily receive incoming calls and do not subscribe to offers, resulting in limited revenue. A telecom operator looking to increase revenue growth wanted to identify these customers and shift them to minimum monthly plans.

Since multivariate analysis had to be performed across features like data subscription purchases, data usage, voice usage, and voice subscription packages, some hybrid machine learning model was designed as the fitting anomaly detection method. The model identified those customers whose behavior did not align with the chosen features. This included customers with no subscriptions and low outgoing usage.

Identifying anomalous data — low-paying customers — helped the client make some strategic decisions. Looking at the outliers’ geography, the operator accordingly revised its subscription plans and adapted its business model.

Achieved results

Just as normal data highlights key patterns, anomalous data detects critical business incidents that may warrant immediate or strategic action. Anomaly detection is an important capability for data-driven organizations. The implemented model, based mainly in unsupervised machine learning algorithms, was a new and effective method to detect anomalies within sample data using multivariate analysis. It was a viable solution for telecom operators looking to understand how to increase revenue, improve efficiency, and reduce cost.

Must have skills of a Data scientist

  • Advanced use and application of data modelling concepts
  • Strong math skills (e.g. statistics, algebra)
  • Proficient in the use of SQL and Python and R
  • Proficient in the use of Excel and Tableau or Power BI
  • Experience working in a cloud environment
  • Business analyst skills to translate business needs into technical requirements
  • Analytics mind-set skill with the ability to turn a business problems into an analytics plan
  • Pragmatic thinking & high degree of attention to detail
  • Experience in machine learning and understanding of the end to end Machine Learning (ML) environment
  • Good communication and stakeholder management skills, including presentation skills, to communicate clearly, regularly, effectively and meaningfully at the senior level of the organization
  • Experience using Git and GitHub

Responsibilities of a Data Scientist

  • Create Machine Learning models and analytics that optimise the customer products
  • Build predictive models and machine-learning algorithms and prepare them for deployment into our live production environment
  • Conduct research and development activities, data exploration and discovery, develop prototypes, algorithms and proof of concepts, using leading data science and innovative Big Data solutions
  • Put forward opportunities that can be solved through advanced analytics and work to develop these
  • Propose solutions and strategies to business challenges
  • Extract valuable business insights and provide recommendations based on those insights
  • Present your findings and recommendations in a visually appealing way to senior stakeholders
  • Undertake pre-processing of structured and unstructured data for your analytics
  • Analyse large amounts of information to discover trends and patterns
  • Contribute to the design of solutions of monitoring and reporting for our models

Challenges of the role

  • Explain technical concepts to non-technical audiences
  • Spend a lot of time with raw data
  • Be flexible and consider context
  • Regular maintenance and version control
  • Understanding the business