Is biostatistics for me

Service navigation

Internship support

You are planning an internship with a statistical reference, e.g. B. as part of the "Case Studies II" event, but you don't know how to pay the additional rent? Additional travel costs prevent you from taking up an internship outside of North Rhine-Westphalia? Discovered great opportunity to write a thesis, but unfortunately not nearby, i. H. over 1.5 hours drive away?

 

The Alumni Association of Dortmund Statisticians e.V. offers students of the Statistics Faculty the opportunity to receive financial support for an external, remote internship.

With 100 € monthly internship support for up to six months, you can at least take some worries away.

How it works? Send an informal application with a description of the internship and your course of study as well as the internship certificate by email to [email protected] The application should be made before the internship begins. Following the internship, we expect a short report, which will be published on this website, for example.

The non-profit alumni association Dortmunder Statisticians e.V. is financed exclusively through the contributions of its members. The internship support is intended to cover costs that are incurred as a result of the internship and that are not covered by the internship provider (e.g. through adequate payment). There is no legal claim. This offer is not binding.

 

We are happy to answer questions about this offer by email.

 

We have already funded these internships:

Kaya Miah visited the German Cancer Research Center in Heidelberg, where she dealt with imputation procedures, among other things.

Laura Zieger examined the results of surveys on working conditions and job satisfaction among teachers in London with regard to their international comparability.

Barbara Brune analyzed gold deposits in Golden (Colorado, USA) using geostatistical methods.

Marc Hüsch dealt with the wind energy forecast in Oldenburg (Germany).

 

 

Testimonials

Report on my external block internship at the German Cancer Research Center in Heidelberg

by Kaya Miah

 

Dear members of the Alumni Association for Dortmund Statisticians e.V.,

For the period from February 19, 2018 to March 29, 2018, I was allowed to complete an external block internship as part of the Case Studies II event in the Biostatistics department of the German Cancer Research Center (DKFZ) in Heidelberg. With more than 3000 employees, 1300 scientists and 400 doctoral students in over 90 departments, the DKFZ is one of the largest biomedical research institutions in Germany. The Biostatistics department at the DKFZ is assigned to the research focus on cancer risk factors and prevention.

 

During my internship, the focus of my day-to-day work was the processing of a statistical project. This field of activity included the investigation of a model-based approach for the multiple imputation of missing realizations of covariates in interaction questions in the case of a Cox regression model as an analysis model. For this purpose, I should analyze the quality of the modified SMC-FCS ("Substantive Model Compatible Fully Conditional Specification") algorithm from the R package of the same name in a simulation study in comparison to various proven imputation methods for variable transformations in the form of interactions. The adapted SMC-FCS method represents a modification of the established regression-based FCS approach, in which missing values ​​are conditionally imputed on the other covariables of examination units with observed realizations on the basis of a correspondingly specified imputation model. For the modified FCS algorithm, the compatibility of the analysis model used for the evaluation with the choice of the imputation model is an essential characteristic of the modeling. The second sub-task then consisted of evaluating the applicability of the various regression-based imputation methods to real data on the basis of an observational study on lymphatic and myeloid diseases, which was carried out at the University Hospital in Heidelberg until 2013.

 

In addition to working on the project, I was also allowed to sit in on statistical consultations with various cooperation partners. Biostatistics collaborates among other things. with surrounding facilities such as the Amyloidosis Center at Heidelberg University Hospital and the National Center for Tumor Diseases (NCT) to support the statistical planning and evaluation of clinical and epidemiological studies.

 

At the end of the internship, the department employees went to the 64th Biometric Colloquium in Frankfurt am Main for three days last week. At this annual meeting of the German Region of the International Biometric Society, I was given the opportunity to participate in lectures on a wide variety of topics in biometrics, among others. from the areas of survival time analysis, bioinformatics, Bayesian statistics or meta-analysis to participate. In addition, I was able to gain an insight into current developments in biometric research.

 

Overall, I enjoyed the internship in Heidelberg very much, as I was able to gain a comprehensive insight into the professional activity of a biostatistician and was able to work with the great team of biostatistics at the DKFZ. Last but not least, I would also like to thank you very much for the support provided by the Alumni Association for Dortmund Statisticians e.V., which provided me with tremendous help during my stay in Heidelberg.

 

Testimonials

Report on my research internship in London

by Laura Zieger

 

Dear members of the alumni association,

From October 2017 to January 2018 inclusive, I did a voluntary internship in London at the non-profit company Education Datalab, which consists of a multidisciplinary group of experts in quantitative educational research. The main task of the company is to carry out independent, relevant research which can be used by politics and schools to improve the current system. These analyzes are based on various national and international large-scale data sets. In addition, Education Datalab is entrusted with the national project management for the OECD study TALIS in Great Britain.

During my internship I dealt with data from a cycle of the TALIS study and analyzed it within two different projects with regard to working conditions and job satisfaction of teachers in an international context. TALIS 2013 surveyed over 100,000 teachers from 37 countries on aspects and views in the various areas of their work, such as technical training, intrinsic attitudes towards teaching and the characteristics of their workplace. The aim of my work was to check which countries can be validly compared with regard to the constructs mentioned above. Even if it is to be expected that, for example, the measurement of job satisfaction in Japan and England shows different things and scores cannot be compared, it is still important to prove this and also to check "similar" countries. For this purpose, the various working conditions and job satisfaction were first identified using methodology from the area of ​​structural equation models. Subsequently, an attempt was made to find the largest possible number of countries that have meaningfully comparable constructs, which was done by considering different levels of measurement invariance. Countries where this is the case can be analyzed accordingly and a ranking can be created. This information can be used by countries, for example, to compare themselves with corresponding other countries and to identify needs and opportunities for action.

I was involved in every step of the project and gained experience there. This begins with a thorough literature research and detailed elaboration of the research question and goes through the data acquisition and the actual analyzes to the summary and the writing down. Articles have emerged from both projects, which will hopefully be published in the relevant specialist journals. As part of these activities, I have acquired or deepened various skills, especially in relation to scientific and independent work.

Overall, I really enjoyed my time in London in many ways. On the one hand, it was super interesting to get an insight into the working life of this area of ​​statistics and to work with a great multidisciplinary team. On the other hand, it was great to live in London and to make long-term contacts. My internship helped me advance personally and “professionally” and I wouldn't want to miss this experience.

London offers many opportunities and is one of the most exciting cities in terms of culture, science and business. Unfortunately, this is also reflected in the cost of living in London, which is why I was very happy to have received additional funding from the Alumni Association of Dortmund Statisticians. Thanks to the monthly financial support, I was able to cover the cost of public transport that I used on my daily commute from my shared apartment to my work place in the heart of London. I would like to thank you very much for that!

 


Report on my research internship in the USA

by Barbara Brune

 

From January to June 2017, I did a research internship as part of Case Studies II in the Department of Mining Engineering at the Colorado School of Mines, Golden, Colorado in the USA. Golden is a small town about 30 miles west of Denver.

 

The Colorado School of Mines is one of the top three universities in the United States for engineering, especially mining and geology. During my research stay, I dealt intensively with geostatistics, an area of ​​spatial statistics that is particularly concerned with estimating resources. Here I have edited and evaluated a data set with drill holes from a gold mine in Northern California to produce a geostatistical

Carry out analysis of the deposit. After a few weeks of familiarizing myself with the theory, I soon started working on my data set. A major hurdle was the software I was supposed to use, a mine planning program that followed all of the procedures

Black box methodology performs: you press a button and results appear, but you have no idea what happened. For me, who is used to programming most things myself in R, it took a lot of getting used to and one or the other time I got used to

driven to despair. But once I accepted that, it was a lot of fun, and it was great when the results came out too!

 

I also attended a total of three lectures at the university, one in the statistics department and two in the mining department. This gave me a very interesting insight into the university system in the USA, which is quite different from the German - it's much more school-based, there is compulsory attendance everywhere and you get mountains of homework that is graded.

 

Overall, I enjoyed my stay in the USA very much. It was very exciting to get an insight into the field of geostatistics. What I was able to learn here is innovative and absolutely relevant to practice. I've been assured umpteen times that the mining companies are looking for you

tear if you know your way around this field. Working with engineers and geologists was a lot of fun, and it was very interesting to see how differently the different disciplines approach problems and how much one can learn from each other through interdisciplinary work.

 

At the very international Colorado School of Mines I have many people from many

got to know different countries that have contributed significantly to

I really enjoyed my stay here. Colorado is a beautiful state with many possibilities and a breathtakingly beautiful landscape!

 

However, I was surprised that the cost of living in Colorado is almost twice as high as in Dortmund: the rent alone in mine - not expensive! - Flat share was higher than

my entire monthly budget in Dortmund! Without the financial support from the Alumni Association, I would have skidded quite a bit, and I would like to take this opportunity to thank you once again for the fact that the Alumni Association made this great time possible for me.

 


Report about my research stay at the Center for Wind Energy Research in Oldenburg

by Marc Hüsch

Dear members of the alumni association,

In the period from May to October 2016, I wrote my master's thesis as part of a research stay at the Center for Wind Energy Research (ForWind) in Oldenburg. Fortunately, thanks to monthly financial support from the Alumni Association, some of the costs that arose from an additional apartment and frequent trips between Dortmund and Oldenburg could be covered. As I am very grateful to the alumni association for their financial support, I would like to give the members of the association a little insight into my research stay in Oldenburg with this report.

The aim of my master's thesis was to investigate the influence of the forecast horizon on spatial dependencies of wind power forecast errors. With regard to the forecast uncertainty, it is known that this increases with increasing forecast horizons. However, it has so far been less investigated that the spatial correlation of prediction errors should also increase due to an increased influence of the analysis error. In large-scale regions with strongly positively correlated errors, an additional risk arises. Since load flow calculations for predicted wind power in Europe must also be high-performance for long forecast horizons (e.g. to calculate cross-border electricity transports), this additional risk factor should be taken into account in the calculations in the future. For this reason, detailed studies of spatial dependencies on prediction errors are of high relevance for many decision-makers.

Wind speeds from numerical weather forecasts from the European Center for Medium-Range Weather Forecasts (ECMWF) for a high-resolution European grid for the period from April 2010 to February 2016 were available for the analysis. 32 European countries, in which a high number of wind power is currently installed, were selected for the analysis. The wind speeds for all grid cells in the study region were initially converted into wind power using a regional wind power curve. For six different forecast horizons (12, 24, 36, 48, 60 and 72 hours) it was then shown with the help of methods from spatial statistics that a longer forecast horizon leads to significantly higher spatial dependencies on forecast errors. A cluster analysis was then used to show spatial differences with regard to the dependencies. It became clear that strong spatial dependencies occur particularly in flat areas in Northern Europe. Since a relatively high number of wind power is installed in these areas, future consideration of the dependencies on wind power forecast errors is of great relevance. In the final part of the analysis, the influence of various influencing variables on the spatial correlation of prediction errors was quantified with the help of classification and regression trees.

 

Overall, I really enjoyed my time in Oldenburg, as I was able to gain a lot of interdisciplinary insights into active research and also make many useful contacts. Further collaborations with ForWind are planned for the future. In particular, interested students from Dortmund should be given the opportunity to do an internship (for example as part of the Case Studies II event) at ForWind or to write a bachelor thesis in the "Energy Meteorology" working group.

Since I already enjoyed the scientific work during my master's thesis, I am looking forward to being able to gain further experience in research as a research assistant at the Institute for Economic and Social Statistics at TU Dortmund University. I would like to thank the alumni association very much for the support they have given and I will remain a member of the association in the future.