Dashboards - Visualising Everyday Life
An Example from Compound Management
Written by Sabrina Leuenberger and Christian Zart
Summary
How does day-to-day compound management work? Does the feeling match the reality? Collecting data and visualizing it in dashboards can provide new views of everyday life. What about processes, communication and monitoring? Dashboards can be used to gain important insights into compound management. But how do you create meaningful dashboards? In this article, we would like to introduce the approach.
Compound Management
Compound Management (CM) manages large libraries of chemical compounds, usually in research institutions. The main task of CM is the storage of compounds and their provision in various forms or concentrations to support the research teams. Easy and efficient access to a wide variety of different compounds is essential, particularly for experiments and screenings in drug research, chemical biology or other scientific fields. To ensure this, processes are required that often involve various departments.
The compound management processes result in a lot of data. By visualising this data in a dashboard, the efficiency of processes and workflows can be better assessed and bottlenecks in particular can be identified.
In addition to internal and cross-departmental process optimisation in Compound Management, the dashboards also make another valuable contribution. A dashboard provides analyses, results and statistics for communication with employees in research or management.
wega has already been able to offer support in various projects for the creation of meaningful dashboards. We would like to share the experience gained from these projects with you here.
Dashboards and KPIs
Dashboards are not just visualised statistics. When set up correctly, they can do much more. The charts and tables can be designed dynamically and various filter options allow you to delve deep into the details. This allows you to create different data view levels.
In addition, KPIs can be defined and made accessible for different interest groups based on the same initial data.
You often "know" where there is a problem in the process or there are complaints that this or that is not working properly in the process. In some cases, we were able to confirm these assumptions based on dashboards and substantiate or refute them with figures. Instead, other areas that have potential for optimisation can suddenly come into focus.
How Do You Get Meaningful Dashboards?
It all starts with a business analysis to understand the process. This also includes the availability of associated data. Many questions need to be answered, such as
- What data is collected per process stage?
- Which time stamps or audit trail entries are available?
- Which time stamps are most suitable?
- What is the data situation, i.e. is reliable data available? Which time calculation between which time stamps makes sense?
- Is there a rule or logic for selecting the correct timestamp from several timestamps in each case?
Process Visualisation
The description and visualisation of the process flow with associated and defined time stamps helps to define the objectives of the dashboard and the KPIs. This alone can bring added value to our customers. Becoming aware of the process can provide initial indications of how and where it could be improved.
An example of the creation of a logic: For time stamps 2 to 4, a logic is used to define which of these time stamps is used in the time calculation. In order for a compound to be ordered, it must be registered and physically present in the warehouse. A standard solution is then produced from this, which can then be ordered. Sometimes the order is created before the solution is available. In this case, it makes sense to start the time calculation only when processing of the order can begin.
It also often makes no sense to display the time required between all time stamps. This means from timestamp 1 to timestamp 2, from timestamp 2 to timestamp 3 etc... For example, it may be more interesting to define the time between the compound registration (timestamp 1) and the start of the first order (timestamp 6) as a KPI.
Nevertheless, the other time expenditures should also be available in order to be able to delve further into the details if required.
Figure 1: Process from receipt of order to provision of compound
Merging the Required Data
Once the desired data has been defined, it is compiled in the required format. This can make up a large part of the project if, for example, the data has to be compiled from different sources. Depending on the complexity, the data may have to be merged in several stages. This is very individual for each project and is also directly related to the technology used or the systems and applications used.
Finally, the data provided is visualised. Microsoft Power BI or Tableau, for example, can be used for this. It is important to use suitable chart types and not to make them too complex. This is the only way to ensure that the diagrams remain readable and meaningful.
Dashboard - Visualising the Data
A practical example where visualisation with the dashboard could help.
A typical compound workflow starts with registration and ends with the completion of one or more analyses. There are many intermediate stages in between, as can be seen in the example of the visualized workflow in Figure 1. A central part is the provision of the compound (time stamp 6) after receipt of an order (time stamp 5).
The time spent between these two time stamps was divided into categories:
- < = 12h ( light green )
- > 12h - < = 24h ( dark green )
- 24h - < = 48h ( yellow )
- > 48h ( red )
The number of orders per category and month is shown.
On the left you can still see large yellow and red bars, which represent a high time expenditure. The process has been optimised over the months, which means that the green bars now dominate on the right.
Figure 2: Time required from receipt of order to provision of compound
This is one possible representation of the collected data. There are other ways of analysing the data. For example, you can change the timeline to days or years. Or to display the orders behind a single red bar. You can also display a table with selected columns in addition to the diagram. Many variations are available.
Findings from the Process Analysis and the Dashboard Graphics
Both our customer and we were able to gain a wide range of insights from the project presented:
- An analysis of the process in question can reveal differences between the actual process and the originally planned process.
- Comparison of the perceived bottlenecks with real data.
- There are process steps that take more time than expected. For example, the compound ordered is provided by Compound Management and is ready for collection. From the CM's point of view, the process is now complete. However, the compound was not collected by the ordering laboratory until days later. From the laboratory team's point of view, their process only starts after the compound has been in the laboratory. To whom is the waiting time now allocated as this is of course part of the time measurement of the overall process? This could be an approach to possibly check the laboratory capacity.
- Temporary overruns in average process times could be reconciled with large quantities of new compounds, such as the purchase of compound libraries.
- Regular monitoring can be carried out using dashboards. This allows trends to be recognised and measures to be taken in good time.
- Easy-to-read diagrams should be created. This includes the number of data sources and the chart type. We can recommend the following article: https://www.perceptualedge.com/articles/visual_business_intelligence/save_the_pies_for_dessert.pdf
Finally, we would like to thank our customers for their cooperation. The projects were a lot of fun and allowed us to contribute our experience. At the same time, we were also able to gain new experiences and insights. Because it should not be forgotten: Every project is individual and unique.