All past updates




  • Our paper “Predicting Subjective Recovery from Lower Limb Surgery Using Consumer Wearables” is published. We showed that passively collected wearable PGHD can capture post-surgery physical activity changes relative to individual’s baseline, and baseline data can improve prediction of self-reported recovery time at 4 weeks post surgery.

  • During the upcoming workshop “The Future of Digital Health”, I will give a talk highlighting “Predicting subjective recovery from lower limb surgery using consumer wearables” paper I co-authored. This workshop will highlight key papers from the newly released Special Issue of Digital Biomarkers, sponsored by Evidation Health and DiMe. Register for the workshop here.

  • During summer of 2020, I am working as a Data Science Intern in Digital Measures team @ Evidation Health. My main project aims at estimating medical procedure recovery trajectories and predicting recovery time from wearable patient-generated health data.

  • Why R? Foundation awards a supporting grant for Women in Data Science to aid exceptional female data scientists in Poland. I serve as a member of Scientific committee for this initiative.


  • I passed my preliminary schoolwide oral exam. My Committee consisted of three faculty from the Dept. of Biostatistics, one faculty from Dept. of Epidemiology, and one faculty from the Dept. of Medicine.

  • I received Leadership, Empowerment and Learning Culture Award during Novartis US Analytics Conference 2019. My conference talk presented the work done as a Sensor Data Analytics Intern with Novartis in Basel, Switzerland during summer 2019.

  • Our paper “Adaptive empirical pattern transformation (ADEPT) with application to walking stride segmentation” just got published in Biostatistics with the journal’s highest reproducibility status! See ADEPT article online.

  • Our R package adept that implements ADEPT pattern-segmentation method is out on CRAN (CRAN index. The vignette shows example of strides segmentation from raw accelerometry data of 25min outdoor run w/ walking and resting bouts. Listed May 2019 top 40 new CRAN packages.

  • I am working as Sensor Data Analytics Intern with Novartis pharmaceutical company in Basel, Switzerland this summer. I am also presenting at ICAMPAM 2019 conference on Jun 26 in Maastricht, the Netherlands.

  • Our R package runstats that provides methods for fast computation of running sample statistics for time series is out on CRAN (CRAN index). Use runstats to compute (1) mean, (2) standard deviation, and (3) variance over a fixed-length window of time-series, (4) correlation, (5) covariance, and (6) Euclidean distance (L2 norm) between short-time pattern and time-series. Implemented methods utilize Convolution Theorem to compute convolutions via Fast Fourier Transform (FFT).

  • Our recent article “Accelerometry Data in Health Research: Challenges and Opportunities” is now published and available online.


  • Interested in statistical methodology for singal processing in physical activity research? Come to JSM session on Wednesday (8/1/2018), 2:00 PM - 3:50 PM.

  • WhyR? 2018 conference starts on July 2nd in Wroclaw, Poland. Both academia and industry professionals meet and discuss experiences in R software development and analysis applications.

  • I am working as a Research Assistant with Drs. Jacek Urbanek and Ciprian Crainiceanu this summer at a novel approach for identifying individual working strides from subsecond accelerometry data of walking.

Marta Karas
Marta Karas
Postdoctoral researcher