Applying business intelligence and machine learning capabilities in the laboratory can not only improve lab performance but can also benefit your business. The latest version of the Data Analytics Solution available with Thermo Scientific™ SampleManager LIMS™ software facilitates deeper levels of management, operation and scientific insight through the data contained within their SampleManager LIMS. The solution delivers insight through two main capabilities – business intelligence and machine learning.
Real-time insights through business intelligence
Business intelligence dashboards address the common business questions confronting analytical laboratories. The business intelligence dashboards in the Data Analytics Solution were created in collaboration with the Thermo Fisher Scientific’s Pharma Services business. These dashboards offer important insight into laboratory activities in real-time and without the need for additional software or data export. You can use these dashboards to make more informed business decisions. For example, dashboards can help you identify whether retests are an issue in your labs, or track stock levels across all your laboratories. The available dashboards include:
- Job Status
- Sample Status
- Job Backlog
- Sample Backlog
- Sample Times
- Location Result Mapping
- Performance Dashboard
- Project Status
- Resource Management Dashboard
- Sample Point Results
- Stock Overview Dashboard
- Laboratory Compliance
Machine learning and Exploratory Data Analysis
The Exploratory Data Analysis capability of the Data Analytics Solution provides a straightforward way to explore the results of the analysis performed on a set of samples. Exploratory data analysis is the fundamental process to start investigations on a data set by discovering patterns and relationships, reviewing the variables’ distributions, outlier identification, and others with the help of summary statistics and graphical representations.
The EDA capability provides the following four tools:
- Univariate Analysis
- Bivariate Analysis
- Linear Regression
- Principal Component Analysis
Predict future resource requirements with machine learning
The Machine Learning – Forecasting capabilities of the Data Analytics Solution can help your business predict resource requirements based on historical data. With a historical data sample and some knowledge of future work, the solution applies machine learning techniques to predict the forthcoming received samples for both the current month and the next one. From the forecast results, the Forecasting capability also calculates estimates for the expected tests, instrument category runs, and stock usage for the current and upcoming month. Anticipating the samples received, the analyses performed, the instruments used, and the stocks consumed gives can help your business:
- Hire/train/relocate personnel to meet future needs.
- Replenish or stop buying related stocks and consumables.
- Restructure the lab according to the incoming type of samples.
- Define and update goals such as future sales, workflow durations and others.
Predict test results with profiling and regression
The Machine Learning – Profiling capability of the Data Analytics solution provides an innovative way to predict the result of a yet-to-be-executed test using previous data and novel machine learning techniques. To accomplish this task, SampleManager LIMS uses the results of previous tests applied to similar samples or jobs as inputs to create a machine learning pipeline that applies the predictive power of various machine learning algorithms such as the gradient tree boosting library called XGBoost, neural networks, random forest, and linear and logistic regression. The capability supports the training and prediction of numerical variables. The user may select a “Calculated,” “Boolean” or “Numerical” test as its target variable. Anticipating the result of a test without conducting it can provide several business benefits including:
- Identifying and eliminating the need to perform redundant tests.
- Reducing the number of samples tested.
- Reducing the usage of expensive reagents and consumables.
- Generating a prediction to all eligible samples without any extra cost or time.
- Taking actions by anticipating the result of the test.
- Failing early samples that would not pass the test.
The Machine Learning Capability provides a no-code, easy-to-use framework for everyone and enables the user to explore data contained in SampleManager LIMS with novel machine learning techniques, evaluate the results and create insightful predictions and analyses.
Anticipate test results and plan ahead
By profiling with time-based data, the Data Analytics Solution can help you predict the result of a time series sample before testing is complete. The capability predictions can now consider a time component for the variables. Measurements done to the same source but at different time steps give valuable information to the model that can be used to predict a specific variable at a given time step. With time based predictions, users can fast-forward time to make proactive decisions and:
- Anticipate final-day results.
- Stop or modify the process if the last-day outcome is not close to the desired one.
Tools like the Data Analytics Solution in SampleManager LIMS software help businesses benefit from more in-depth insight into their laboratory data. Want to learn more about the data visualization capabilities available in SampleManager LIMS software? Watch our advancing data visualization in the lab webinar on-demand or visit our Data Analytics Solution website
Leave a Reply