Benchling is the leading cloud platform for biotech R&D, helping scientists plan studies, collaborate, and capture structured experimental data. Prior to this work, users could carry out experimental analysis through Insights Dashboards. Dashboards allow users to write SQL queries to query Benchling data, then visualize that data as charts. Customers currently use Insights Dashboards for a variety of use cases, including scientific analytics (what were the results of this assay?) and operational analytics (how many notebook entries are currently in review?)
Scientists need to analyze raw experimental data in order to make meaningful sense of it. In June 2024, I took a break from molecular biology and worked with the analysis team to build Insights Analysis. The Analysis tool allows users to visualize, interact and transform experimental data iteratively.
Raw data outputted from lab instruments is messy. Sometimes all it takes is a quick tidy: filtering, reshaping, outlier removal. Sometimes scientists have to do a deep clean: curve fitting, harmonizing, doing linear regressions. Benchling allows scientists to do both via SQL queries that extract existing data. The flexibility of SQL makes Insights a very powerful tool for technical users, but difficult to use for many lab scientists who are not comfortable writing code.
Optimizing for lab scientist use cases
The lab scientist's responsibilities include: performing experimental procedures, analyzing results of experiments through statistical analyses, and generating reports for their manager to review findings. They ask experiment specific questions such as: What were the results of this assay? Are there any unusual outliers in this data? Do I need to re-run this assay? These questions can often be answered through simple raw data transformations and basic data visualizations.
If a scientist lacks the ability to write SQL (which, unsurprisingly, many do) they often analyze raw experimental data using tools they are familiar with such as Excel, JMP, or Prism. This disruption often leads to many scientists not bringing analysis results back into Benchling, leaving experiments only partially captured within the platform.
For IT teams, in order to manage this problem, they either have to incentivize scientists to learn SQL or continue to maintain a distributed software stack to accommodate preferred point-solution analysis tools.
Insights Analysis empowers users to answer scientific questions without leaving Benchling, improving the efficiency and quality of workflows, reducing the time spent on manual tasks like data wrangling and confirming data quality, and reducing the chances of errors when shuttling data from one tool to another.
Benchling's Insights platform offers robust tools for data analysis, enabling users to create, manipulate, and visualize datasets derived from various sources within the Benchling ecosystem. The key considerations that influenced the designs were:
