Search

This paper proposes a novel multimodal deep weakly-supervised learning framework, SinkholeNet, to classify and localize sinkhole(s) in high-resolution RGB-slope aerial images. The SinkholeNet first employs a multimodal Convolutional Neural Network (CNN) architecture that simultaneously extracts features from the input RGB image and ground slope map and then fuses the extracted features. It then uses an improved ShuffleNet architecture on the fused features to classify patches as sinkholes or non-sinkholes. Finally, the last extracted feature maps, belonging to the sinkhole class, are used as input of gradient-weighted class activation mapping (Grad-CAM) to localize sinkhole(s) in a weakly-supervised setting. The proposed weakly-supervised framework intends to increase the available labeled data for training and decrease the cost of human annotation. We also introduce a novel publicly available weakly labeled sinkhole dataset comprising RGB-slope paired image patches to support reproducible research. The experimental results on the newly introduced dataset show that the SinkholeNet outperforms the other methods considered in this paper both for sinkhole classification and localization. © 2023 National Authority of Remote Sensing & Space Science

Besides providing electron density profiles (EDP), GNSS Radio Occultation (GNSS-RO) measurements allow monitoring the frequency and the areas where ionospheric scintillations occur. In this work, RO measurements composing an experimental data set are processed with the back propagation (BP) method to estimate the location of sporadic E-clouds and equatorial plasma bubbles (EPB). The data set includes non-conventional measurements tracked up to 600 km (generally around 80 km), covering F-region heights, shortly before MetOp-A was decommissioned. Results indicate the combination of extended occultations and the BP method is promising for monitoring the occurrence and characterizing ionospheric irregularities in the F-region and the E-region. © 2023 International Union of Radio Science., Radio Occultation Accuracy for Climate, Meteorology, and Space Weather - Part 2

The growing availability of high-dimensional data sets offers behavioral scientists an unprecedented opportunity to integrate the information hidden in the novel types of data (e.g., genetic data, social media data, and GPS tracks, etc.,) and thereby obtain a more detailed and comprehensive view towards their research questions. In the context of clustering, analyzing the large volume of variables could potentially result in an accurate estimation or a novel discovery of underlying subgroups. However, a unique challenge is that the high-dimensional data sets likely involve a significant amount of irrelevant variables. These irrelevant variables do not contribute to the separation of clusters and they may mask cluster partitions. The current paper addresses this challenge by introducing a new clustering algorithm, called Cardinality K-means or CKM, and by proposing a novel model selection strategy. CKM is able to perform simultaneous clustering and variable selection with high stability. In two simulation studies and an empirical demonstration with genetic data, CKM consistently outperformed competing methods in terms of recovering cluster partitions and identifying signaling variables. Meanwhile, our novel model selection strategy determines the number of clusters based on a subset of variables that are most likely to be signaling variables. Through a simulation study, this strategy was found to result in a more accurate estimation of the number of clusters compared to the conventional strategy that utilizes the full set of variables. Our proposed CKM algorithm, together with the novel model selection strategy, has been implemented in a freely accessible R package.

This paper is focused on multiple supervised machine learning algorithms' performance evaluation in terms of classification accuracy and speed for the blind frequency bands classification into three occupancy classes: white, gray, and black spaces for potential implementation in cognitive radio application. Training and validation data sets consisting of 180 000 samples, including 60 000 samples per class, have been collected in the controlled experiment. Data samples have been generated using a hardware signal generator and recorded on the receiver's front end as the time-domain complex signals. Gray space data samples contain one, two, or three signals modulated into 2FSK, BPSK, or QPSK with symbol rates 10, 100, or 1000 kSymbol/s. White space data samples contain no own generated signals. Black space data samples contain two signals with the symbol rate of 22.5 MSymbol/s and offset +14 MHz and -14 MHz from the central frequency occupying the entire observation band. Training and validation of twenty supervised machine learning algorithms have been performed offline in the Matlab Classification Learner application using the collected data set. Fine decision trees have demonstrated the highest classification accuracy of 87.8 %, the observed classification speed of 630000 Objects/s is also higher than the required 2000 Objects/s. Medium decision trees and ensemble boosted trees have demonstrated 87.5 % and 87.7 % accuracy and classification speeds of 950000 and 230000 Objects/s respectively. Therefore, ensemble boosted trees, and fine and medium decision trees have been selected for the deployment on the target radio application in the scope of future work.

Requirements quality literature abounds with publications presenting artifacts, such as data sets and tools. However, recent systematic studies show that more than 80% of these artifacts have become unavailable or were never made public, limiting reproducibility and reusability. In this work, we report on an attempt to recover those artifacts. To that end, we requested corresponding authors of unavailable artifacts to recover and disclose them according to open science principles. Our results, based on 19 answers from 35 authors (54% response rate), include an assessment of the availability of requirements quality artifacts and a breakdown of authors’ reasons for their continued unavailability. Overall, we improved the availability of seven data sets and seven implementations. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

Interactive visualization is an essential tool for data analysis. Cloud-based data analysis software must handle growing data sets without relying on powerful end-user hardware. This thesis explores and tests various methods to speed up primarily time series plots of large data sets on the web for the biotechnology research company Sartorius. To increase rendering speed, I focused on two main approaches: downsampling and hardware acceleration. To find which sampling algorithms suit Sartorius's needs, I implemented multiple alternatives and compared them quantitatively and qualitatively. The results show that downsampling increases or eliminates data set size limits and that test users favored algorithms maintaining local outliers. With hardware acceleration that substantially increased the amount of simultaneously rendered points for more detailed representations, these methods pave the way for efficient visualization of large data sets on the web.

The dataset is produced within the framework of the HORIZON 2020 project called MEDIADELCOM (Critical Exploration of Media Related Risks and Opportunities for Deliberative Communication: Development Scenarios of the European Media Landscape) in 2021-2022. The dataset is one of the 14 single-country data sets included in the consolidated file of country data sets (with 5623 entries), all in msw.xlsx format. All tables are searchable by 20 variables: full reference, year of publication, nationall/international publication, language, country the publication deals with, time of empirical data gathering, type of publication, open access/not OA, where referenced, focus on journalism domain, focus on media-related competences domain, focus on media usage patterns domain, focus on legal and ethical regulations domain, type of the approach, original key words, main topic, comments, country. As the data has been gathered specifically about the research done in four mentioned domains concerning potential ROs emanating from the news media development for deliberative communication, this database does NOT cover ALL the academic publications in the fields of media and journalism research. Consequently, the above-mentioned conditions limit the generalizations and comparisons based on the current database., Mediadelcom

The freshwater raphidophyte Gonyostomum semen forms extensive summer blooms in northern European humic lakes. The development of these blooms might be facilitated by a lack of natural top-down control, as few zooplankton species are able to prey on these large algal cells (up to 100µm) that expel trichocysts upon physical stress. In this study, we describe a small ciliate species (<17µm) that preys on G. semen by damaging the cell membrane until cytoplasm and organelles spill out. Sequencing of clonal cultures of the ciliate tentatively identified it as the prostomatid species Urotricha pseudofurcata. Grazing experiments illustrated that feeding by U. cf. pseudofurcata can significantly reduce cell concentrations of the microalga. However, differences in cell size and growth rate between two investigated ciliate strains resulted in noticeably different grazing pressure. Environmental sequencing data from five different lakes supported potential interactions between the two species. Urotricha cf. pseudofurcata might, thus, play an important role in aquatic ecosystems that are regularly dominated by G. semen, reducing the abundance of this bloom-forming microalga and enabling transfer of organic carbon to higher trophic levels., A pre-print of the associated manuscript is available on ResearchSquare https://assets.researchsquare.com/files/rs-3251412/v1/83927ab2-6709-4101-a55f-c0b99c6d619c.pdf?c=1692283936

The group presented here is the Jadgal people. Similar to other groups in this region, such as the Afro-Baloch, they are not aware of their origin. They consider themselves to be separate from the Baloch and call themselves Jadgal. Likewise, Baloch people also consider them to be a separate ethnic group. Jadgal people are classified by tribal affiliation. Jadgali speakers are multilingual, typically speaking several or all of Persian, Balochi, Jadgali, Urdu and Sindhi. Only their language distinguishes them from others in the region. They share the same traditions and customs with the Baloch people., The collection is an outcome of an ongoing project documenting the Jadgali language spoken in Sistan and Balochistan, Hormozgan and Kerman Provinces in the Islamic Republic of Iran.

The group presented here is the Brahui people. Similar to other groups in this province, such as the Afro-Baloch and Jadgal, they are not aware of their origin. Some consider themselves separate from the Baloch and call themselves Brahui, and others just call themselves Balochi. However, the Baloch people consider them to be a separate ethnic group and call them Brahui. Similar to Baloch people, Brahui people are classified based on their own tribes. Brahui speakers are multilingual, typically speaking several or all of Persian/Sistani Persian, Balochi, and Brahui, and some are good at Pashto. Only their language distinguishes them from other groups in the region. They share the same traditions and customs with the Baloch people. The new generation does not speak Brahui; instead they have switched to either Balochi or Persian.

The group presented here is the Baloch. The Balochi language belongs to the northwest Iranian branch of the Iranian languages, which in turn belong to the Indo-Iranian branch of Indo-European. It has three main dialects: Southern, Eastern and Western Balochi. Each of these dialects presents its own sub-divisions (see Jahani and Korn 2009). Balochi is mostly spoken in southeastern Iran and southwestern Pakistan,but is also found in Afghanistan, Turkmenistan, Oman and the UAE. All Baloch people call themselves Baloch except for Koroshi people, who call themselves either Korosh or Qashqai, although their language is Balochi., In the current version of the collection, the corpus consists of approximately 80 hours of recorded naturally spoken folktales; Fairy tales; Legends; Islamic tales.

In 2017, ticks and field data were collected from 12 different sites in Stockholm County originally chosen as random controls for another study but was never used. In 2019, we collected ticks and field data at 35 randomly selected sites along the natural-urban gradient. To calculate and urbanization index, we used the proportion of artificial surfaces surrounding each site. All sampling sites were visited once with a total of 295 sampling plots inventoried for ticks and field data. For each sampling plot, we recorded date, time, temperature, weather conditions, number of ticks, vegetation height and tree stem density surrounding the inventory plot. To retrieve large landscape characteristics, we established 10 buffer zones ranging from 100m to 1000m around each sampling site in GIS using satellite land cover maps (retrieved from: https://www.naturvardsverket.se/verktyg-och-tjanster/kartor-och-karttjanster/nationella-marktackedata/ladda-ner-nationella-marktackedata/). These maps have a spatial resolution of 10m and include the following main categories 1) Forest and seminatural areas, 2) Open areas, 3) Arable land, 4) Wetlands, 5) Artificial surfaces and 6) Inland and marine water. These main categories are further divided into subcategories with detailed information regarding the different land cover classes. In the analyses, we used the main categories, with the exception of Forest and seminatural areas where we included eight individual forest types: Pine forest, Spruce forest, Mixed coniferous forest, Mixed forest, Broadleaved forest, Broadleaved hardwood forest, Broadleaved forest with hardwood forest and Temporarily non-forest. To calculate landscape configuration metrics at each sampling site, we used land cover data from the GIS buffers with a 1000m radius, exported to GeoTIFF format and analyzed them with FRAGSTATS version 4. For landscape heterogeneity we used Shannons’ diversity index (SHDI) and to measure the aggregation of landscape attributes we used Co, För att förstå utbredningen av den vanliga fästingen Ixodes ricinus, undersökte vi hur både lokala faktorer och olika landskapsegenskaper påverkar förekomsten av fästingar i grönområden längs med urbaniseringsgradienten i Stockholms län, Sverige. Fästingar och fältdata samlades in under 2017 och 2019 och analyserades i relation till habitatdata och landskapsdata framtaget från geografiskt informationssystem (GIS). 2017 samlades fästingar och fältdata in från 12 olika platser i Stockholms län som ursprungligen hade valts som slumpmässiga kontroller till en annan studie men aldrig använts. 2019 samlades fästingar och fältdata in från 35 olika platser längs med urbaniseringsgradienten. Dessa 35 platser valdes slumpmässigt ut med hjälp av GIS. För att beräkna urbaniseringsgraden för varje plats, använde vi andelen exploaterad mark i en 1000m buffertzon. För de 12 första insamlingsplatserna under 2017 valde vi slumpmässigt ut 10 2m × 2m rutor för inventering av fästingar och fältdata. På de 35 platser som vi inventerade 2019, samlade vi in fästingar i 5 2m × 2m rutor. Vi besökte alla platser en gång och totalt inventerades 295 provtagningsrutor. För varje ruta noterade vi datum, tid, väderförhållande, antal fästingar, vegetationshöjd och trädstamstäthet runt varje ruta. För att ta fram landskapsdata skapades 10 buffertzoner från 100m till 1000m radies runt varje insamlingsplats i GIS tillsammans med Naturvårdverkets Nationella Marktäckedata. Dessa satellitkartor har en upplösning på 10m och har följande huvudkategorier 1) Skog, 2) Öppen mark, 3) Åkermark, 4) Öppen våtmark, 5) Exploaterad mark, och 6) Vatten. För att identifiera riskfaktorer valde vi att använda dessa huvudkategorier med undantag for kategorin Skog där vi valde att inkludera de åtta olika skogstyperna: Tallskog, Granskog, Barrblandskog, Lövblandad barrskog, Triviallövskog, Ädellövskog, Triviallövskog med ädellövsinslag och Temporärt ej skog. För att ta fram statistik och beräkna konfigurationen av landska

Uppgifter om migration (såväl lokal som interregional och internationell) i Sverige och Finland under åren 1613–1618. Uppgifterna är på hushålls- eller individnivå och rör såväl landsbygd som städer. För varje hushåll/individ anges (i den mån det uppges i källan) titel/yrke, namn, ursprungsort samt destination för migrationen (angivet med socken/stad, härad och landskap/land). Uppgifterna är hämtade från rannsakningslängderna från Älvsborgs andra lösen (i Riksarkiven i Stockholm och Helsingfors)., Data on migration (local as well as interregional and international) in Sweden and Finland during the period 1613–1618. The data concern the household/individual level, urban and rural. For each household/individual the data contain title/occupation, name, home address, and migration destination; given as parish/town, hundred, and county/country (if the data is present in the source). The data is collected from the taxation records from the second Älvsborg's ransom taxation (in the National Archives of Sweden and Finland).

This datafile contains information about movements at a global scale, that were identified as being composed mainly by mothers, fathers and others who have close connections with youth and children, on behalf of whom they seek and demand change. The data summarized in this file is about groups, movements and organizations made up in large extent of mothers who in that role engage in activism and related activities in the public sphere.

Requirements quality literature abounds with publications presenting artifacts, such as data sets and tools. However, recent systematic studies show that more than 80% of these artifacts have become unavailable or were never made public, limiting reproducibility and reusability. In this work, we report on an attempt to recover those artifacts. To that end, we requested corresponding authors of unavailable artifacts to recover and disclose them according to open science principles. Our results, based on 19 answers from 35 authors (54% response rate), include an assessment of the availability of requirements quality artifacts and a breakdown of authors’ reasons for their continued unavailability. Overall, we improved the availability of seven data sets and seven implementations. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

test

2023-04-06 Version 2.0 published. Minor changes in the structure of the data file. There were no changes in the data.2023-03-23 The title was changed fromSource data for “The effects of water column dissolved oxygen concentrations on lake methane emissions: Results from a whole-lake oxygenation experiment” to "Source data for ”Higher apparent gas transfer velocities for CO2 compared to CH4 in small lakes”"2022-11-30 Version 1.0 published.

Interactive visualization is an essential tool for data analysis. Cloud-based data analysis software must handle growing data sets without relying on powerful end-user hardware. This thesis explores and tests various methods to speed up primarily time series plots of large data sets on the web for the biotechnology research company Sartorius. To increase rendering speed, I focused on two main approaches: downsampling and hardware acceleration. To find which sampling algorithms suit Sartorius's needs, I implemented multiple alternatives and compared them quantitatively and qualitatively. The results show that downsampling increases or eliminates data set size limits and that test users favored algorithms maintaining local outliers. With hardware acceleration that substantially increased the amount of simultaneously rendered points for more detailed representations, these methods pave the way for efficient visualization of large data sets on the web.

A standard classifier is forced to predict the label of every test instance, even when confidence in the predictions is very low. In many scenarios, it would, however, be better to avoid making these predictions, maybe leaving them to a human expert. A classifier with that alternative is referred to as a classifier with reject option. In this paper, we propose an algorithm that, for a particular data set, automatically suggests a number of accuracy levels, which it will be able to meet perfectly, using a classifier with reject option. Since the basis of the suggested algorithm is conformal prediction, it comes with strong validity guarantees. The experimentation, using 25 publicly available two-class data sets, confirms that the algorithm obtains empirical accuracies very close to the requested levels. In addition, in an outright comparison with probabilistic predictors, including models calibrated with Platt scaling, the suggested algorithm clearly outperforms the alternatives.

The dataset includes time-series data collected by four different sensors, which measure two target gases, Hydrogen Sulfide and Methyl Mercaptan, in the presence of air. To obtain measurements, each gas was individually exposed to the multi-sensor setup while in the presence of air. The dataset is particularly useful for gas classification tasks, as deep learning and data fusion techniques can be applied to identify the target gases.

Requirements quality literature abounds with publications presenting artifacts, such as data sets and tools. However, recent systematic studies show that more than 80% of these artifacts have become unavailable or were never made public, limiting reproducibility and reusability. In this work, we report on an attempt to recover those artifacts. To that end, we requested corresponding authors of unavailable artifacts to recover and disclose them according to open science principles. Our results, based on 19 answers from 35 authors (54% response rate), include an assessment of the availability of requirements quality artifacts and a breakdown of authors’ reasons for their continued unavailability. Overall, we improved the availability of seven data sets and seven implementations. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

Requirements quality literature abounds with publications presenting artifacts, such as data sets and tools. However, recent systematic studies show that more than 80% of these artifacts have become unavailable or were never made public, limiting reproducibility and reusability. In this work, we report on an attempt to recover those artifacts. To that end, we requested corresponding authors of unavailable artifacts to recover and disclose them according to open science principles. Our results, based on 19 answers from 35 authors (54% response rate), include an assessment of the availability of requirements quality artifacts and a breakdown of authors’ reasons for their continued unavailability. Overall, we improved the availability of seven data sets and seven implementations. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

Requirements quality literature abounds with publications presenting artifacts, such as data sets and tools. However, recent systematic studies show that more than 80% of these artifacts have become unavailable or were never made public, limiting reproducibility and reusability. In this work, we report on an attempt to recover those artifacts. To that end, we requested corresponding authors of unavailable artifacts to recover and disclose them according to open science principles. Our results, based on 19 answers from 35 authors (54% response rate), include an assessment of the availability of requirements quality artifacts and a breakdown of authors’ reasons for their continued unavailability. Overall, we improved the availability of seven data sets and seven implementations. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

Requirements quality literature abounds with publications presenting artifacts, such as data sets and tools. However, recent systematic studies show that more than 80% of these artifacts have become unavailable or were never made public, limiting reproducibility and reusability. In this work, we report on an attempt to recover those artifacts. To that end, we requested corresponding authors of unavailable artifacts to recover and disclose them according to open science principles. Our results, based on 19 answers from 35 authors (54% response rate), include an assessment of the availability of requirements quality artifacts and a breakdown of authors’ reasons for their continued unavailability. Overall, we improved the availability of seven data sets and seven implementations. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

The material consists of exam results among 458 students who for the first time answered a certain written exam on the last of four modules in an academic first-cycle (undergraduate) course in history at a Swedish university, distributed across ten course instances over eleven years, 2010-2020. The exam results consist of a score divided into points received on answers to the written exam and on points that each student was able to obtain from optional hand-in assignments before the written exam (quizzes). The material also contains information about the exam results from the two immediately preceding modules on the same course and at the same course instance. The data have been used to examine the extent to which the quizzes have affected the result on the final written exam and the extent to which the outcome has been affected by the examination result on the two immediately preceding modules.

This dataset contains the configurations files used and logs produced for my bachelor degree project in Informatics. Logs are produced output from the Kubernetes disk benchmarking tool dbench: https://github.com/leeliu/dbench Logs can be found in the 'a20filan_bachelors_work_logs.zip' zip file and subdivided in folders by platform, storage solution, and test occasion. Configuration files for deploying test sequence, and setting up infrastructure and storage solutions, can be found in 'a20filan_bachelors_work_configuration.zip' and is subdivided by platform, storage solution or infrastructure. Folder 'AWS' contains the configuration files for deploying infrastructure and setting up storage solutions and test procedure on Amazon Elastic Kubernetes Service Infrastructure for AWS can be found in folder 'aws/learn-teraform-provision-eks-cluster' and is based on terraform script from Hashicorp: https://developer.hashicorp.com/terraform/tutorials/kubernetes/aks Folder 'azure' contains the configuration files for deploying infrastructure and setting up storage solutions and test procedure on Microsoft Azure Kubernetes Service Infrastructure for Azure can be found in folder 'azure/learn-teraform-provision-aks-cluster' and is based on terraform script from Hashicorp: https://developer.hashicorp.com/terraform/tutorials/kubernetes/eks

This includes all data from the project "Phytoplankton blooms and pCO2" that has been used and referred to in the manuscript "Carbon dioxide reduction by photosynthesis undetectable even during phytoplankton blooms". Raw data generated by Karla Münzner is found in the folder "Raw data", data used in statistical analysis in the folder "R data". The folder "Protocols" holds all lab protocols used for dna analyses (DNA extraction, DNA quantification and qPCR). There is also a pdf ("Lab notes dna analyses") that records the day-to-day work of all dna-related laboratory work conducted during the project.

Context: Code smells are patterns in source code associated with an increased defect rate and a higher maintenance effort than usual, but without a clear definition. Code smells are often detected using rules hard-coded in detection tools. Such rules are often set arbitrarily or derived from data sets tagged by reviewers without the necessary industrial know-how. Conclusions from studying such data sets may be unreliable or even harmful, since algorithms may achieve higher values of performance metrics on them than on models tagged by experts, despite not being industrially useful. Objective: Our goal is to investigate the performance of various machine learning algorithms for automated code smell detection trained on code smell data set(MLCQ) derived from actively developed and industry-relevant projects and reviews performed by experienced software developers. Method: We assign the severity of the smell to the code sample according to a consensus between the severities assigned by the reviewers, use the Matthews Correlation Coefficient (MCC) as our main performance metric to account for the entire confusion matrix, and compare the median value to account for non-normal distributions of performance. We compare 6720 models built using eight machine learning techniques. The entire process is automated and reproducible. Results: Performance of compared techniques depends heavily on analyzed smell. The median value of our performance metric for the best algorithm was 0.81 for Long Method, 0.31 for Feature Envy, 0.51 for Blob, and 0.57 for Data Class. Conclusions: Random Forest and Flexible Discriminant Analysis performed the best overall, but in most cases the performance difference between them and the median algorithm was no more than 10% of the latter. The performance results were stable over multiple iterations. Although the F-score omits one quadrant of the confusion matrix (and thus may differ from MCC), in code smell detection, the actual differences are minimal

When bacterial communities mix, immigration history can fundamentally affect the community composition as a result of priority effects. Priority effects arise when an early immigrant exhausts resources and/or alters habitat conditions, thereby influencing the establishment success of the late arriver. The strength of priority effects is context dependent and expected to be stronger if environmental conditions favor the growth of the first arriver. In this study we conducted a two-factorial experiment testing the importance of nutrient availability and grazing on the strength of priority effects in complex aquatic bacterial communities. We did so by mixing two dissimilar communities, simultaneously, and with a 38 hours’ time-delay. Priority effects were measured as the invasion resistance of the first community to the invading second community. We found stronger priority effects in treatments with high nutrient availability and in absence of grazing, but in general the arrival timing was less important than the selection by nutrients and grazing. At the population level, the results were complex, but priority effects may have been driven by bacteria belonging to for example the genera Rhodoferax and Herbaspirillum. Our study highlights the importance of arrival timing in complex bacterial communities especially in environmental conditions which favor rapid community growth.

This package contains items for 9 real-world software projects. The data is supposed to aid the detection of the presence of the Fire Drill anti-pattern. We include data, ground truth, code, and notebooks. The data supports two distinct methods of detecting the AP: a) through issue-tracking data, and b) through the underlying source code. Therefore, this package includes the following: Original data: For each project, its original artifacts (e.g., wikis, meeting minutes, mentor's notes, etc.) Evaluation of raters' notes by the assessor Fire Drill in issue-tracking data: Ground truth for whether and how strong each project exhibits the Fire Drill AP, on a scale from [0,10]. This was determined by two individual raters, who also reached a consensus. Coefficients for indicators for the first method, per project. Detailed issue-tracing data for each project: what occurred and when. Time logs for each project. Fire Drill in source-code data: Four technical reports that document the developed method of how to translate a description into a detectable pattern, and to use the pattern to detect the presence and to score it (similar to the rating). Also includes a report for how activities were assigned to individual commits. Source code density data (metrics) for each commit in each of the nine projects as a separate dataset. Code: a snapshot of the repository that holds all code, models, notebooks, and pre-computed results, for utmost reproducibility (the code is written in R).