APR Nov/Dec 2022 - 43

« MANUFACTURING
Figure 4. Screenshots of the DT Dashboard on the PTDC-F homepage. A: Embedding of the individual dashboard pages into the
department website. B: Interactive dashboard plot showing the percentage of different features (tasks) in individual epics (projects) in the
'data consumption' initiative. The table below the pie chart displays links to the corresponding tools. C: Status of different features for the
epic 'PI Server Integration', also interactively linked with the table below.
and 11-33% respectively). Due to this diversity, various software had
to be utilized (Figure 5A). We concluded that generalists with strong
knowledge in R™ and Python™ programming, but also visualization,
data handling and problem-solving skills
positions best. In terms of their philosophy,15
suited our embedded
the tools developed were
mostly used for descriptive analysis (42-67% of projects per year), but
query and predictive analysis also constituted many projects every
year (17-26% and 11-21%). The focus on intermediate complexity of
interpretation reflects that DT is neither at its start nor completely
mature in our department. The projects added value in many
different ways (Figure 5B). The main benefits21
are productivity, team
collaboration and traceability, which together make up 50-67% of the
total project efforts each year.
Over the years, the percentage of projects realized specifically for the
Sciences section housing the interns strongly decreased (Figure 5C).
Correspondingly, there was a continuous increase in the percentage
of projects on department (PTDC-F) or division wide scale (PTDC),
showing the success of the activity. Since the tools were developed
for a broader community, there were higher user numbers per project
in the last year. Where in the first two years all projects were designed
for <50 people, in the following year 32% of the tools were intended
for 50-99 users and even 15% for >100 people. The time savings per
user were quite low for the first year's projects, since it included much
valuable prototyping. These savings increased in the following years
since tools for more concrete use cases were developed.
Altogether, the three generations of interns successfully delivered
a manifold of value to a broad spectrum of people. The results
demonstrate the diversity of opportunities available for DT tools in
our Pharmaceutical R&D environment, and the increasing department
interest in data science activities.
Figure 5. Portfolio analysis of projects led by the interns (V.L., K.L., J.P.). A: Software usage in projects per year. B: Quality improvement in
projects per year. C: Collaboration in projects per year.
www.americanpharmaceuticalreview.com |
| 43
»

http://www.americanpharmaceuticalreview.com

APR Nov/Dec 2022

Table of Contents for the Digital Edition of APR Nov/Dec 2022