Client Profile: A medium-sized, not-for-profit media organisation specialising in audiovisual content distribution.
Challenge: The client needed to enhance their understanding of user behavior and effectively segregate bot traffic from human interactions to minimise cloud data infrastructure costs.
Solution: We provided a contemporary solution that refined the analysis of download logs, enabling the client to gain deeper insights into their audience demographics and behaviours. Our solution also provided a significant reduction in the time required to perform the analysis and reporting and handled additional edge cases.
Impact:
- Cost Efficiency: Streamlined data analysis helped in identifying cost-saving opportunities, reducing unnecessary expenditure on cloud resources.
- Staff Empowerment: We conducted flexible staff mentoring sessions, equipping the staff with modern analytics tools, which one staff member described as “light years ahead” of their previous capabilities. Also up-skilled in the use of LLM-powered development environment which was readily embraced and adopted.
- Collaborative Platform: Developed a lower-code web application to demonstrate the possibilities for rapid distribution of analytic capability within their organisation.
- Innovative Technology: Assisted in the evaluation and implementation of cutting-edge language detection algorithms using leading AI platforms like Hugging Face.
Benefits:
- Enhanced Productivity: The right selection of analytics tools boosted the productivity of the internal team.
- Infrastructure Cost Reduction: Smart data analysis practices led to more efficient and timely use of cloud services.
- Data-Driven Insights: Clearer understanding and segmentation of their download traffic.
This case study exemplifies how our analytics consultancy can transform data management and analysis, leading to cost savings and improved operational efficiency. It is a testament to our commitment to providing state-of-the-art solutions, typically only accessible to larger organisations, that cater to the dynamic needs of today’s media organisations.
Our project leveraged a robust and scalable technical stack, all mature open source components, designed to meet the demands of modern data analytics and user interface design.
- Client Infrastructure: Google Cloud Services (GCS), PostgreSQL
- Data Sources:
- Log Files (GCS): Analysed detailed logs to extract meaningful patterns and insights.
- Audio: Processed audio content for input into language models.
- Analytics Tools:
- DuckDB: An analytics database optimised for fast and in-depth data analysis.
- Pandas: A powerful library for data manipulation and analysis, providing high-performance data structures.
- Jupyter: Interactive notebooks that offer a dynamic environment for live code, visualisations, and narrative text.
- Quarto: A publication tool that brings data analysis to life with rich, reproducible documents.
- Web app framework:
- Streamlit: An open-source app framework that turns data scripts into shareable web apps with ease.
- Modeling Frameworks:
- SpeechBrain: An open-source and all-in-one speech toolkit for speech analysis and processing.
- Hugging Face: Leveraged the leading AI research platform to evaluate state-of-the-art language detection algorithms.
The integration of these open source technologies facilitated a seamless workflow from data collection to insight generation, empowering the client to make data-driven decisions with confidence.