SFL Scientific is answering real business challenges for our clients
through innovation and deep expertise in data science.
Donor contributions are a significant part of success for the American Heart Association (AHA). SFL Scientific performed unsupervised machine learning techniques to segment its customer base, profile existing donors, and discover characteristics of high-value donors and their utilization. SFL was able to unlock highly specific demographic and geographic correlations which helped the AHA to build campaigns, assign, and manage resources for maximum efficiency.
Read more about our solutions in Advertising & Marketing >
American Well was seeking help with patient volume forecasting for planning in their telemedicine services. American Well provided SFL Scientific with historic, state-level, daily patient volume data from the past three years. SFL chose to use a combination of ARIMA and TBATS models to analyze the time-series data provided. The model outperformed the previously generated forecasts. As a pilot to showcase how to data science can optimize operations, lowering the Mean Absolute Percentage Error (MAPE) derived from the SFL model significantly inefficiency and improved resource planning.
Anheuser-Busch (AB) wanted to leverage data that they have collected on its Mexico operations to drive revenue, so AB contacted SFL Scientific to segment their customer base to better understand customer behavior. SFL created unsupervised, hierarchical clustering models to determine high-value customers and overall spending patterns for AB's entire Mexican operation. Through our machine learning and cluster analysis, AB was able to create target marketing campaigns and strategies to drive sales and maximize overall profitability.
Read more about our customer segmentation & customer 360 services >
Biodesix develops non-invasive, blood-based and liquid biopsy diagnostic tests for oncology. SFL Scientific was given a very limited patient dataset comprised of Dynamic Light Scattering (DLS) frequency spectrum readings. SFL implemented a time series classification algorithm to predict cancer from the DLS readings and achieved a 97% F-measure score against an independent, blinded, hold-out dataset. SFL also created an instrumentation and cross-tool calibration model that replaced outdated development tools during development.
SFL Scientific developed a set of machine learning models to determine the top attributes that are most deterministic of winning a client deal. By modeling the sales process, personnel, and examining operational data across the entire sales organization we can produce greater insight, monitor pitfalls, and provide feedback to leaders and employees. Our work enabled the Bullhorn to optimize their sales process and maximize sales.
Read more about our solutions in Business Intelligence >
C+A Global is involved in the manufacturing and distribution of consumer products and photographic equipment. SFL developed a deep learning model with real-time object recognition capabilities for live video streams to integrate with smart security camera systems. The smart security system, powered by the deep learning model, was hosted on AWS GPU instances and monitors video streams from security cameras, classifying hundreds of objects and scenarios in the real-time. The system can be optimized for event processing and facial recognition, creating alarms for various scenarios and automatically notifying appropriate authorities when emergency situations are identified.
Read more about our solutions in Consumer Electronics >
Cunesoft wished to use natural language processing to extract key information from hundreds of thousands of medical documents. The documents were written in ten different languages and ranged from Electronic Health Records (EHR) to drug pamphlets. To solve this problem, SFL created an ensemble model of feature-based class predictions and several sequence labeling methods- HMMs, CRFs, n-grams, etc. Through Microsoft Azure, SFL Scientific wrapped the model in a web API for easy access, upload, and management as well as future scaling.
SFL worked with Cybric to add machine intelligence on top of their state-of-the-art orchestration and cloud-based Security-as-a-Service platform. We developed a data science platform to predict vulnerabilities in various cybersecurity-related scenarios.
Explore how machine learning can enable enhanced Cybersecurity >
Farmers Edge is the leader in precision agriculture. They provide advanced agronomic solutions to identify and map field variability, optimizing crop inputs that result in higher yields, better quality, and less environmental impact. FarmersEdge wished to provide customers with crop forecasting. SFL Scientific built a forecasting solution across millions of acres by employing deep learning models and state-of-the-art pixel intensity convolutional neural networks in conjunction with standard growth and vegetative index features to predict the field-level crop yields. SFL is an expert in working with big image data and developing solutions for GIS and satellite image analysis applications.
Explore our solutions for AgTech and Farming >
LinkSquares approached SFL Scientific about devising an automated solution to algorithmically extract key terms from contracts and legal documents for their cloud platform. SFL provided LinkSquares with an NLP algorithm comprised of a stacking model ensemble using XGBoost as the meta-classifier to extract key terms from each legal document and classify tokenized text into pre-defined categories. The system SFL delivered was then deployed through AWS for scalability and on-demand access.
Read our case study with Linksquares >
L’Oréal's R&D and product innovation division wished to develop a software framework to classify acne severity using user-uploaded photographs. SFL Scientific developed a sophisticated classification solution via convolutional neural networks to tackle the inherent variability in the skin and image data, as well as to handle real-world constraints. SFL obtained high match accuracy for predicting the acne severity when comparing to trained dermatologists. As acne is not only a physical condition, where studies have linked acne to emotional conditions such as depression, anxiety, and low self-esteem, the ability to recommend treatments and engage customers is profound. By using an image recognition algorithm to diagnose the severity of acne, L’Oréal can extend their recommendation engine, create solutions for quick diagnosis, and aid their clients without access to a dermatologist in making better decisions.
SFL Scientific performed customer and market segmentation analysis for a large Japanese retailer. The dataset consisted of millions of historic purchases for tens of thousands of clients. SFL examined dozens of features in the transcations, inventory, customer profile, user behavior, and log data. Segments were created using a hierarchical clustering algorithm and a deep profile was created for each. This was a first step in building more complex models around customer churn, promotions, incentives, and loyalty programs.
SFL Scientific is a consulting partner working for Optum on helping them innovative, automate, and draw insights from their healthcare, medical, and big data tables and records. Utilizing our expertise in creating end-to-end pipelines around natural language processing, Optum continues to leverage data science and analytics in increasing profitability, health outcomes, disease detection, and improved patient journey through a number of initiatives.
SFL Scientific works with Quanterix by providing technical and professional services for selecting and optimizing filtering, detection, and classification algorithms for their flagship product. Our embedded effort is focused on maintaining and tuning existing systems while providing development for upgrading future systems involved in determining the concentration of biomaterials and other biological compounds through image processing.
Read more about our solutions in the Life Sciences >
In order to promote a corporate culture of fairness and equality, Staples sought SFL Scientific’s help to identify intra-company wage gaps based on gender and ethnicity. SFL selected a base case demographic and modeled a wage predictor based on this demographic. SFL then applied this model to predict wages for all employees of Staples to identify any wage gaps present. Through this effort, Staples was able to understand the significant contributors to wage gaps and made key business decisions from this effort as part of their “Employers for Pay Equity” Consortium Initiatives.
The Danish energy company, SE, suspected significant under-reporting of energy usage which was negatively impacting revenue. SE Energy & Climate, as the climate division of SE, one of Denmark's largest utility companies and a modern energy and communications supplier wanted to forecast energy demand at the regional meter and sub-meter level. In order to solve this problem, SFL Scientific combined dynamic time warping (DTW) and anomaly detection algorithms to forecast demand, segment user device usage, and ultimately generate a probability of fraud/misreporting for each customer based only on their time-stamped energy meter data.
Read more about our solutions in Energy & Utilities >
SFL worked with Salesforce’s HR department to accurately map employees and their skills. We created a topic model of text-based employee data that corresponded to well understood skillsets. By mixing in some semi-supervised learning, we developed a probabilistic modeling of how all employees would fit in every job role, thus allowing for automatic internal transfer recommendations.