In today’s data-driven world, businesses must efficiently process and analyze vast amounts of data to remain competitive. High-efficiency big data solutions are necessary for swiftly and accurately transforming raw data into actionable insights. One such innovative solution provider is TIMi, a company at the forefront of big data analytics. TIMi provides tools designed to handle large datasets with minimal infrastructure.
The Genesis of TIMi
TIMi was founded in Ath, Belgium, on August 3, 2007, by Frank Vanden Berghen. The acronym TIMi stands for “The Intelligent Mining Machine,” a tribute to the technologies embedded in it, such as Artificial Intelligence (AI), Machine Learning (ML), and Data Mining. The name also nods to an old MS-DOS game, “The Incredible Machine,” reflecting the founders’ innovative spirit.
Evolution of TIMi Software
TIMi’s journey began with developing the TIMi Modeler, a commercially available automated machine learning (Auto-ML) tool. The initial version of Modeler debuted in early 2009 and quickly showcased its capabilities by ranking 18th in the prestigious 2009 KDD Cup, outperforming over 1,200 competitors. This achievement underscored TIMi Modeler’s status as a leading Auto-ML tool.
The software suite expanded in June 2010 with Stardust, a clustering tool designed to handle large datasets commonly found in sectors like banking, telecom, insurance, and retail. Stardust employs advanced algorithms, such as K-Means++ and PCA, to create accurate clusters for large populations, enhancing customer segmentation and targeting.
In January 2011, TIMi introduced Anatella, a revolutionary data transformation tool focused on analytical tasks. Unlike traditional ETL (Extract-Transform-Load) tools, Anatella was designed to streamline the preparation of datasets for AI and ML algorithms. Its efficiency and speed have made it the core component of the TIMi software suite.
Global Expansion and Technological Milestones
TIMi’s international footprint began to grow in 2014 with the opening of its first subsidiary outside Europe, TIMi Latin Americas, in Bogota, Colombia. This expansion continued in March 2015 through a partnership with Real Impact, extending TIMi’s reach across Africa. TIMi’s efficient engine, coded in C and Assembler, enabled it to process vast amounts of data on minimal infrastructure, a critical advantage for telecom operators in Africa. For example,TIMi’s technology enables operators to analyze Call Data Records (CDR) from a large number of subscribers daily, helping to develop predictive models for churn and cross-selling with a high degree of accuracy.
In May 2016, TIMi integrated open-source technologies into its suite, allowing users to leverage R, Python, and Hadoop for data transformations and predictive modeling. This integration brought additional functionalities, such as charting and some additional common predictive algorithms like XGBoost, enhancing the versatility of TIMi’s tools.
The software suite expanded further in December 2016 with the launch of Kibella, a self-service Business Intelligence (BI) tool. Kibella enables users to create interactive dashboards quickly and easily, using only a web browser. Its open-source nature ensures no dependencies on third-party tools, providing flexibility and control over data visualization.
Cutting-Edge Features and Industry Leadership
TIMi has continually advanced data analytics technology. In August 2017, Stardust introduced a Virtual Reality (VR) mode, making TIMi the first machine learning tool available in VR. This feature allows users to explore large datasets in a 3D virtual environment, similar to the interface seen in the film “Minority Report.”
In 2018, TIMi exhibited the speed of its data transformation engine, Anatella, using the TPC-H benchmark. This vendor-neutral benchmark involves running 22 SQL queries on a database, and TIMi’s performance was remarkable. The results showed that a single TIMi-based server could outperform a Spark-based cluster composed of several hundred servers, highlighting TIMi’s efficiency and processing power.
Another direct example of TIMi’s high efficiency is illustrated by the work of the company “Snype” in Tanzania. There, Snype is tasked with recalculating the exact net revenue of all telecom companies in the country, ensuring equitable taxation of these same revenues. These computations start from the granular level of data: the raw Call-Data-Records (CDRs) from every telecom subscriber in Tanzania.
Processing such a large volume of data, including information from 120 million SIM cards and a primary CDR table growing by 10 billion rows daily, presents a significant challenge. Remarkably, all these computations are managed on a single machine using TIMi, without requiring the extensive server infrastructure that other solutions might need. This highlights TIMi’s ability to efficiently handle complex and data-intensive operations.
TIMi’s big data solution includes a suite of tools designed to manage large datasets with exceptional speed and precision. From the Auto-ML capabilities of TIMi Modeler to the data transformation features of Anatella and the self-service BI offered by Kibella, TIMi provides a comprehensive platform for data analytics. By continually integrating advanced technologies and expanding its global footprint, TIMi stays at the forefront of big data solutions, supporting businesses in maximizing the value of their data.
Published by: Khy Talara
