In OD van november 2019 publiceerde ik een reactie-column op de column van Daan van Beek aangaande het ‘data lake’. Ik noemde deze reactie Een dure les!, omdat, zoals zo vaak bij dit soort nieuwe technologie, organisaties, bang om achter te blijven, met beide voeten vooruit erin springen om er achteraf achter te komen dat dat misschien toch niet zo handig was. Ondertussen is er een hele hoop geld in gestopt, maar is er wel heel veel ‘geleerd’. Tenminste dat mag gehoopt worden. Het bijgaande bestand bevat zowel de column van Daan van Beek als mijn reactie daarop.
DOCUMENT
Learning is all about feedback. Runners, for example, use apps like the RunKeeper. Research shows that apps like that enhance engagement and results. And people think it is fun. The essence being that the behavior of the runner is tracked and communicated back to the runner in a dashboard. We wondered if you can reach the same positive effect if you had a dashboard for Study-behaviour. For students. And what should you measure, track and communicate? We wondered if we could translate the Quantified Self Movement into a Quantified Student. So, together with students, professors and companies we started designing & building Quantified Student Apps. Apps that were measuring all kinds of study-behaviour related data. Things like Time On Campus, Time Online, Sleep, Exercise, Galvanic Skin Response, Study Results and so on. We developed tools to create study – information and prototyped the Apps with groups of student. At the same time we created a Big Data Lake and did a lot of Privacy research. The Big Difference between the Quantified Student Program and Learning Analytics is that we only present the data to the student. It is his/her data! It is his/her decision to act on it or not. The Quantified Student Apps are designed as a Big Mother never a Big Brother. The project has just started. But we already designed, created and learned a lot. 1. We designed and build for groups of prototypes for Study behavior Apps: a. Apps that measure sleep & exercise and compare it to study results, like MyRhytm; b. Apps that measure study hours and compare it to study results, like Nomi; c. Apps that measure group behavior and signal problems, like Groupmotion; d. Apps that measure on campus time and compare it with peers, like workhorse; 2. We researched student fysics to see if we could find his personal Cup-A-Soup-Moment (meaning, can we find by looking at his/her biometrics when the concentration levels dip?); 3. We created a Big Data lake with student data and Open Data and are looking for correlation and causality there. We already found some interesting patterns. In doing so we learned a lot. We learned it is often hard to acquire the right data. It is hard to create and App or a solution that is presenting the data in the right way and presents it in a form of actionable information. We learned that health trackers are still very inprecise. We learned about (and solved some) challenges surrounding privacy. Next year (2017) we will scale the most promising prototype, measure the effects, start a new researchproject and continu working on our data lake. Things will be interesting, and we will blog about it on www.quantifiedstudent.nl.
LINK
Machine learning models have proven to be reliable methods in classification tasks. However, little research has been done on classifying dwelling characteristics based on smart meter & weather data before. Gaining insights into dwelling characteristics can be helpful to create/improve the policies for creating new dwellings at NZEB standard. This paper compares the different machine learning algorithms and the methods used to correctly implement the models. These methods include the data pre-processing, model validation and evaluation. Smart meter data was provided by Groene Mient, which was used to train several machine learning algorithms. The models that were generated by the algorithms were compared on their performance. The results showed that Recurrent Neural Network (RNN) 2performed the best with 96% of accuracy. Cross Validation was used to validate the models, where 80% of the data was used for training purposes and 20% was used for testing purposes. Evaluation metrices were used to produce classification reports, which can indicate which of the models work the best for this specific problem. The models were programmed in Python.
DOCUMENT
Wanneer we over HRM en technologie spreken, kunnen we niet meer heen om HR analytics. Gefaciliteerd door de alsmaar groeiende hoeveelheid beschikbare data, oftewel Big Data, proberen organisaties momenteel volop waardevolle inzichten uit de bijna oneindige hoeveelheid data te genereren. Samenwerking tussen wetenschap en praktijk ligt voor de hand. De één kan goed analyseren, de ander beschikt over een schat aan data. Toch komen samenwerkingsverbanden vaak niet verder dan het inzetten van een afstudeerder of het verzorgen van een workshop. Anders gezegd: er wordt volop gedate en er vinden veel one-night stands plaats, maar tot duurzame relaties komt het vaak niet. Waarom niet? En hoe zouden we de samenwerking dan wel vorm kunnen geven?
DOCUMENT
Machine learning models have proven to be reliable methods in classification tasks. However, little research has been conducted on the classification of dwelling characteristics based on smart meter and weather data before. Gaining insights into dwelling characteristics, which comprise of the type of heating system used, the number of inhabitants, and the number of solar panels installed, can be helpful in creating or improving the policies to create new dwellings at nearly zero-energy standard. This paper compares different supervised machine learning algorithms, namely Logistic Regression, Support Vector Machine, K-Nearest Neighbor, and Long-short term memory, and methods used to correctly implement these algorithms. These methods include data pre-processing, model validation, and evaluation. Smart meter data, which was used to train several machine learning algorithms, was provided by Groene Mient. The models that were generated by the algorithms were compared on their performance. The results showed that the Long-short term memory performed the best with 96% accuracy. Cross Validation was used to validate the models, where 80% of the data was used for training purposes and 20% was used for testing purposes. Evaluation metrics were used to produce classification reports, which indicates that the Long-short term memory outperforms the compared models on the evaluation metrics for this specific problem.
DOCUMENT
Aanleiding : Het vakgebied Customer Experience is de laatste jaren enorm in ontwikkeling. Organisaties zien de toegevoegde waarde van een positieve klantbeleving. Voor commerciële organisaties kan een positieve klantbeleving leiden tot meer tevreden klanten die loyaler zijn naar de organisatie, meer bereid zijn de organisatie aan te bevelen (NPS) en minder gevoelig zijn voor prijs. Voor organisaties in de publieke sector kan een goede klantbeleving daarnaast leiden tot een beter imago en meer vertrouwen in de organisatie. Omdat klantbeleving een steeds belangrijker plek inneemt op de agenda van organisaties heeft het lectoraat Marketing en Customer Experience in 2021 besloten om een onderzoek te doen naar de toekomst van het vakgebied customer experience. Het belangrijkste doel van dit onderzoek was om duidelijk te krijgen hoe deze toekomst er mogelijk uit komt te zien en wat hiervan uiteindelijk de consequenties zijn voor nader onderzoek, onderwijs en de beroepspraktijk.
LINK
De publieke sector is in digitale transformatie en biedt haar diensten in toenemende mate digitaal. Tegelijkertijd hebben publieke dienstverleners te maken met de uitdaging om de dienstverlening 'persoonlijk' bij individuele burgers aan te laten sluiten. Hogeschool Utrecht deed de afgelopen twee jaar samen met Belastingdienst, Vitens, RET en een aantal andere publieke organisaties onderzoek naar hoe die twee zaken te combineren. In deze whitepaper bieden wij op basis van een groeimodel vier aanbevelingen die publieke organisaties helpen hun dienstverlening succesvol te digitaliseren en tegelijk persoonlijk te laten aansluiten bij hun burgers.
LINK
From the article: Abstract Sub-chronic toxicity studies of 163 non-genotoxic chemicals were evaluated in order to predict the tumour outcome of 24-month rat carcinogenicity studies obtained from the EFSA and ToxRef databases. Hundred eleven of the 148 chemicals that did not induce putative preneoplastic lesions in the sub-chronic study also did not induce tumours in the carcinogenicity study (True Negatives). Cellular hypertrophy appeared to be an unreliable predictor of carcinogenicity. The negative predictivity, the measure of the compounds evaluated that did not show any putative preneoplastic lesion in de sub-chronic studies and were negative in the carcinogenicity studies, was 75%, whereas the sensitivity, a measure of the sub-chronic study to predict a positive carcinogenicity outcome was only 5%. The specificity, the accuracy of the sub-chronic study to correctly identify non-carcinogens was 90%. When the chemicals which induced tumours generally considered not relevant for humans (33 out of 37 False Negatives) are classified as True Negatives, the negative predictivity amounts to 97%. Overall, the results of this retrospective study support the concept that chemicals showing no histopathological risk factors for neoplasia in a sub-chronic study in rats may be considered non-carcinogenic and do not require further testing in a carcinogenicity study.
DOCUMENT
Abstract Background: Frail older adults who are hospitalized, are more likely to experience missed nursing care (MNC) due to high care needs, communication problems, and complexity of nursing care. We conducted a qualitative study to examine the factors affecting MNC among hospitalized frail older adults in the medical units. Methods: This qualitative study was carried using the conventional content analysis approach in three teaching hospitals. Semi-structured interviews were conducted with 17 nurses through purposive and snowball sampling. The inclusion criteria for the nurses were: at least two years of clinical work experience on a medical ward, caring for frail older people in hospital and willingness to participate. Data were analyzed in accordance with the process described by Graneheim and Lundman. In addition, trustworthiness of the study was assessed using the criteria proposed by Lincoln and Guba. Results: In general, 20 interviews were conducted with nurses. A total of 1320 primary codes were extracted, which were classified into two main categories: MNC aggravating and moderating factors. Factors such as “age-unfriendly structure,” “inefficient care,” and “frailty of older adults” could increase the risk of MNC. In addition, factors such as “support capabilities” and “ethical and legal requirements” will moderate MNC. Conclusions: Hospitalized frail older adults are more at risk of MNC due to high care needs, communication problems, and nursing care complexity. Nursing managers can take practical steps to improve the quality of care by addressing the aggravating and moderating factors of MNC. In addition, nurses with a humanistic perspective who understand the multidimensional problems of frail older adults and pay attention to their weakness in expressing needs, can create a better experience for them in the hospital and improve patient safety.
DOCUMENT
Social and Emotional Learning programs, designed to enhance adolescents’ social and emotional skills, are implemented in schools worldwide. One of these programs is Skills4Life (S4L), for students in Dutch secondary education. To strengthen this program and adapt it to students’ needs, we conducted an exploratory study on their perspectives on their own social-emotional development, focusing on low-achieving students in prevocational education. We interviewed eleven boys and eleven girls in five focus groups on (1) their general school life experiences, (2) their perceptions and experiences regarding interactions with peers, the problems they encountered in these interactions, and (3) the strategies and skills they used to solve these problems. Driven by findings in related studies initial thematic analyzes were extended using a three-step approach: an inductive, data-driven process of open coding; axial coding; and selective coding, using the social-emotional skills comprised in an often-used SEL framework as sensitizing concepts. Overall, students were satisfied with their relationships with classmates and teachers and their ability to manage their daily interaction struggles. Their reflections on their interactions indicate that the skills they preferred to use mirror the social-emotional skills taught in many school programs. However, they also indicated that they did not apply these skills in situations they experienced as unsafe and uncontrollable, e.g., bullying and harassment. The insights into adolescents’ social-emotional skills perceptions and the problems they encountered with peers at school presented here can contribute to customizing school-based skills enhancement programs to their needs. Teacher training is required to help teachers gain insight into students’ perspectives and to use this insight to implement SEL programs tailored to their needs.
MULTIFILE
