Preventatively reducing workplace injuries with data-driven insights
Delta Lake, data science, machine learning, ETL, AWS
Many industrial injuries don’t happen overnight. They develop slowly, over long periods of repetitive harmful motion, ultimately leading to physical damage that can, in worst cases, be irreparable. This IT services and consulting firm is tackling this problem with wearable devices that track daily motion and activity as well as the wearer’s immediate surroundings. With Knoldus, they can ingest massive amounts of real-time IoT data for downstream machine learning that provides proprietary safety scores and classifications of activities to predict risk, resulting in a smarter, safer environment. They have reduced industrial workplace injury rates by over half and delivered millions in health and insurance cost savings for both their customers and their customers’ employees.
The complexity of leveraging data to improve workplace safety
Industrial injury is a big problem that can have significant cost implications for the employer and the employees who must bear the brunt of medical costs.
“Lower-back injuries are the most common types of injury in the industrial workspace. Every time a worker gets injured, it typically costs more than $65,000 in medical expenses,” explained Bryant Eadon, CIO.
The company’s goal is to capture every relevant data point—roughly 1.2 million data points per day per person—to predict injuries and prevent these runaway costs from occurring. With such large volumes of time-series data flowing in real-time, they struggled to build reliable and performant ETL pipelines that could scale to meet data science needs. Maintaining infrastructure also required significant resources, often taking an entire week to stable provision clusters to handle their workloads.
From a data science perspective, working from a single laptop proved to limit their ability to perform ad hoc queries efficiently, and they weren’t able to train their models against their entire datasets.
Across the various data teams, collaboration among both systems and personnel was challenging. Data professionals already struggle with collaboration as teams are often siloed. Historically, their jobs were less about cross-pollination and more about urgency, but without the right tools to foster the teamwork needed, it just exacerbated the situation.
A unified data lake and streamlined machine learning lifecycle
With the Knoldus unified data analytics Solution, iteration and collaboration are no longer an issue as data engineering, data science, and analysts can more easily work on the data together.
Delta Lake solved their data reliability issues, allowing them to ingest real-time IOT data from various sources quickly. With data pipelines flowing seamlessly to the data science team, the data science team could more easily innovate with machine learning. MLflow streamlined the entire machine learning lifecycle to ensure the best models make it to production.
“Before Knoldus, I had no way to structure my data science research project. If I had a model and iterated 20 times, I would forget what the results for my first model were, so I would have to dig through so much,” said Siva Bommireddy, Data Scientist. “MLflow makes that easier to manage and solves for the iterative nature of data science in general.”
The analyst team was the last group to benefit from the unification of data across the organization. Matt added that being able to produce results for non-technical teams has been incredibly fulfilling. “I can actually deliver results that make sense to all teams, data-specific or not, within 15 minutes,” he said. “Knoldus has solved so many data use cases.”
The organization has since expanded the use of Knoldus expertise Solution to other projects, including one where Kafka is being used to standardize and move data from Apache Cassandra databases to Molecula’s Cloud Data Access platform. “This solution uses multiple Knoldus’ expertise Solution features,” their team members explained. “We structure the data from our Cassandra databases using a model stored in Schema Registry, and we use Knoldus Replicator to replicate topics across multiple datacenters.”
“We move nearly 1.5 trillion dollars through our platform each year, so reliability is critical for us; we cannot have data loss or message-write failures.” As we continue to extend our platform into loan origination, loan decisioning, and other areas, the need to reliably share data becomes more critical. Having Knoldus’ expertise as part of our software architecture enables us to easily move data across products and across data centers, public and private, to fulfill that need.”
60% reduction in injury resulting in over $5M in cost savings
This IT services and consulting firm can now unlock insights from their sensor data, translating into new strategies their customers can employ to improve workplace safety and the livelihoods of their employees potentially.
After performing a deep dive analysis of one of their largest Fortune 100 customers, they measured a reduction of workplace injury by up to 60%, delivering a 355% ROI on $5,347,368 in gross savings. At the same time, they have reduced the margin of error for evaluating injury risk from 23% to just 5%—an improvement of 78%.
“We are in the business of protecting the Industrial Athlete.,” explained Eadon. “Knoldus allows us to unleash the power of data and machine learning to help workplaces become safer, more productive, and a better environment for tens of thousands of industrial workers that we count on in our own everyday lives.”
Knoldus has greatly improved collaboration within our cross-functional data team, empowering us to collectively work towards new data-driven innovations to improve workplace safety.