Low-quality Data Research Articles

Robust Event Classification Using Imperfect Real-World PMU Data

This paper studies robust event classification using imperfect real-world phasor measurement unit (PMU) data. By analyzing the real-world PMU data, we find it is challenging to directly use this dataset for event classifiers due to the low data quality observed in PMU measurements and event logs. To address these challenges, we develop a novel machine learning framework for training robust event classifiers, which consists of three main steps: data preprocessing, fine-grained event data extraction, and feature engineering. Specifically, the data preprocessing step addresses the data quality issues of PMU measurements (e.g., bad data and missing data); in the fine-grained event data extraction step, a model-free event detection method is developed to accurately localize the events from the inaccurate event timestamps in the event logs; and the feature engineering step constructs the event features based on the patterns of different event types, in order to improve the performance and the interpretability of the event classifiers. Based on the proposed framework, we develop a workflow for event classification using the real-world PMU data streaming into the system in real time. Using the proposed framework, robust event classifiers can be efficiently trained based on many off-the-shelf lightweight machine learning models. Numerical experiments using the real-world dataset from the Western Interconnection of the U.S power transmission grid show that the event classifiers trained under the proposed framework can achieve high classification accuracy while being robust against low-quality data.

Why “Solar Tsunamis” Rarely Leave Their Imprints in the Chromosphere

Solar coronal waves frequently appear as bright disturbances that propagate globally from the eruption center in the solar atmosphere, just like the tsunamis in the ocean on Earth. Theoretically, coronal waves can sweep over the underlying chromosphere and leave an imprint in the form of Moreton wave, due to the enhanced pressure beneath their coronal wave front. Despite the frequent observations of coronal waves, their counterparts in the chromosphere are rarely detected. Why the chromosphere rarely bears the imprints of solar tsunamis remained a mystery since their discovery three decades ago. To resolve this question, all coronal waves and associated Moreton waves in the last decade have been initially surveyed, though the detection of Moreton waves could be hampered by utilizing the low-quality Hα data from the Global Oscillations Network Group. Here, we present eight cases (including five in the Appendix) of the coexistence of coronal and Moreton waves in inclined eruptions where it is argued that the extreme inclination is key to providing an answer to address the question. For all these events, the lowest part of the coronal wave front near the solar surface appears very bright, and the simultaneous disturbances in the solar transition region and the chromosphere predominantly occur beneath the bright segment. Therefore, evidenced by observations, we propose a scenario for the excitation mechanism of the coronal-Moreton waves in highly inclined eruptions, in which the lowest part of a coronal wave can effectively disturb the chromosphere even for a weak (e.g., B-class) solar flare.

Using data mining technology to explore causes of inaccurate reliability data and suggestions for maintenance management

Reliability data reflects equipment safety and provides a reference for setting inspection period, thereby serving as crucial information for the implementation of equipment integrity management policies. The calculation foundation of reliability data is maintenance records of adequate data quality. However, maintenance records of doubtful quality are common. Despite excluding poor quality recodes and using only the remaining maintenance recodes to calculate the reliability data, the calculated results generally lack a sufficient degree of confidence. This study applied data mining technology, including quality metrics, the association rule, and clustering, to explore the cause of low-quality maintenance data. The results revealed that the low data quality of maintenance records was due to ineffective maintenance policies, the low integrity of key system columns, nonadherence to the policy, and misunderstanding of column definitions. The proposed method successfully identified the causes of low-quality maintenance records. By incorporating the method into the function module of a CMMS, operators can equip the system with self-diagnosis, self-supervision, and continuous optimization functions.

Impact of imperfection in medical imaging data on deep learning-based segmentation performance: An experimental study using synthesized data.

Clinical data used to train deep learning models are often not clean data. They can contain imperfections in both the imaging data and the corresponding segmentations. This study investigates the influence of data imperfections on the performance of deep learning models for parotid gland segmentation. This was done in a controlled manner by using synthesized data. The insights this study provides may be used to make deep learning models better and more reliable. The data were synthesized by using the clinical segmentations, creating a pseudo ground-truth in the process. Three kinds of imperfections were simulated: incorrect segmentations, low image contrast, and artifacts in the imaging data. The severity of each imperfection was varied in five levels. Models resulting from training sets from each of the five levels were cross-evaluated with test sets from each of the five levels. Using synthesized data led to almost perfect parotid gland segmentation when no error was added. Lowering the quality of the parotid gland segmentations used for training substantially lowered the model performance. Additionally, lowering the image quality of the training data by decreasing the contrast or introducing artifacts made the resulting models more robust to data containing those respective kinds of data imperfection. This study demonstrated the importance of good-quality segmentations for deep learning training and it shows that using low-quality imaging data for training can enhance the robustness of the resulting models.

Multimodal Biometric Fusion Algorithm Based on Ranking Partition Collision Theory

Score-based multimodal biometric fusion has been shown to be successful in addressing the problem of unimodal techniques’ vulnerability to attack and poor performance in low-quality data. However, difficulties still exist in how to unify the meaning of heterogeneous scores more effectively. Aside from the matching scores themselves, the importance of the ranking information they include has been undervalued in previous studies. This study concentrates on matching scores and their ranking information and suggests the ranking partition collision (RPC) theory from the standpoint of the worth of scores. To meet both forensic and judicial needs, this paper proposes a method that employs a neural network to fuse biometrics at the score level. In addition, this paper constructs a virtual homologous dataset and conducts experiments on it. Experimental results demonstrate that the proposed method achieves an accuracy of 100% in both mAP and Rank1. To show the efficiency of the proposed method in practical applications, this work carries out more experiments utilizing real-world data. The results show that the proposed approach maintains a Rank1 accuracy of 99.2% on the million-scale database. It offers a novel approach to fusion at the score level.

Development and evaluation of Artificial Intelligence techniques for IoT data quality assessment and curation

Nowadays, data is becoming the new fuel for economic wealth and creation of novel and profitable business models. Multitude of technologies are contributing to an abundance of information sources which are already the baseline for multi-millionaire services and applications. Internet of Things (IoT), is probably the most representative one. However, for an economy of data to actually flourish there are still several critical challenges that have to be overcome. Among them, data quality can become an issue when data come from heterogeneous sources or have different formats, standards and scale. Improving data quality is of utmost importance for any domain since data are the basis for any decision-making system and decisions will not be accurate if they are based on inadequate low-quality data. In this paper we are presenting a solution for assessing several quality dimensions of IoT data streams as they are generated. Additionally, the solution described in the paper actually improves the quality of data streams by curating them through the application of Artificial Intelligence techniques. The approach followed in our work has been to append data quality information as metadata linked to each individual piece of curated data. We have leveraged linked-data principles and integrated the developed AI-based IoT data curation mechanisms within a Data Enrichment Toolchain (DET) that employs the NGSI-LD standard to harmonize and enrich heterogeneous data sources. Furthermore, we have evaluated our design under experimental research conditions, achieving a robust compromise between functionality and overhead. Besides, it demonstrates a stable and scalable performance.

Classification-based process parameter recommendation in sheet metal forming

For the production of sheet metal parts for car bodies, continuous adjustment of process parameters is required to maintain the desired part quality in the presence of scattering blank properties. The digital transformation enables data-driven methods for finding process parameters instead of a time-consuming experience-driven trial-and-error approach. However, it is still hard to measure quality for every part due to cost and technical limitations. Removing data points of low-quality parts helps in recommending proper process parameters. In this paper, we propose a classification-based solution enhanced by a particular data preprocessing for recommending cushion cylinder forces. The solution utilizes anomaly detection and knowledge-based methods to remove potential low-quality data points without quality measures in data prepossessing. On the processed data, a classification model is trained to predict process parameters according to blank properties. Our solution detects 32.46% low-quality parts and gives competitive performance (94.30% prediction accuracy) compared to a model trained on data comprising quality measures.

Integrating process dynamics in data-driven models of chemical processing systems

Data-driven models require high-fidelity data of sufficient quantity and granularity. This is challenging in a complex chemical processing system due to frequent sensor breakdown, process shutdown, malfunctioning of equipment, random fluctuations, miscalibration, inconsistent sampling frequencies, and data entry errors. Thus many models scoring well on training data flounder on the real-time data of industrial systems. This work presents a process dynamics-guided neural network (PDNN) model to improve model generalization that can maintain higher performance in sparse and low-quality data. This has been enacted by adding an additional layer in the neural network architecture to incorporate process dynamics such as material and energy balance equations, universal laws, standard correlations, and field knowledge. We evaluated the proposed model against a standard neural network on a regression and a classification tasks representing a steady state and transient behavior of processing systems. The proposed model yielded improved outcomes on reduced sample-sized data and in extrapolated regimes implying a higher generalization capability of the PDNN model. The proposed process dynamics-guided neural network can be employed as a robust model for handling generalization issues of data-driven methods in processing systems.

LongArms: Fraud Prediction in Online Lending Services Using Sparse Knowledge Graph

Gang fraud, the major and primary security issue in online lending services, can be efficiently solved by the data-driven paradigm that is recognized as a promising solution for online lending gang fraud prediction. However, it is challenging that such predictions need to detect evolving and increasingly impalpable fraud patterns based on low-quality data, i.e., very preliminary and coarse applicant information. The technical difficulty mainly stems from two factors: the extreme <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">deficiency of information associations</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">weakness of data labels</i> . In this work, we mainly address the challenges by enhancing the utility of associations (i.e., <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">recovering missing associations</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">mining underlying associations</i> ) on a knowledge graph. Specifically, we first propose an efficient method of Chinese address disambiguation to recover some critical associations that are broken by the ambiguity of applicant information, e.g., address related information. Then, to mine the implicit associations, we design a novel association representation method, called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Adaptive Connected Component Embedding Simplification Scheme</i> (ACCESS), which can adaptively implement embedding for different connected components depending on their sizes. Finally, we adopt the graph clustering algorithms and devised predicting schemes based on the above enhanced associations to predict gang fraud in the case of weakness of data labels. Moreover, we propose a framework called RMCP by integrating the above techniques, which is consists of four steps: <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Recovering</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Mining</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Clustering</i> , and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Predicting</i> , for efficiently predicting gang fraud. The good performance is validated by the experiments on a real-world dataset from a commercial lending company. Meanwhile, we provide a visual decision support system named <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LongArms</i> over the RMCP framework.

In-situ sensor calibration for building HVAC systems with limited information using general regression improved Bayesian inference

Sensors in building heating, ventilation and air-conditioning systems (HVACs) play important roles in maintaining indoor environmental quality and energy consumption. Owing to the repeatedly varied outdoor working environment and indoor users’ demand, sensor faults could be inevitable in the lifespan. To allow HVACs worked at fault-tolerant way, previous studies developed the in-situ sensor calibration method via energy conservation equations and Bayesian inference (EC-BI). However, the practical application may encounter challenges like limited-variable information, low-quality data and increasing risks of calibration uncertainty by indirect information supplement. These cause increasing in-situ calibration complexity and modeling costs. To address these challenges, this study proposed a general regression improved Bayesian inference (BI) in-situ sensor calibration strategy. The multiple linear regression (MLR) was utilized as a typical example of general regression method to improve the BI method. The proposed MLR-BI method was validated using both simulated and practical data of two building HVAC systems in two case studies. The principle component analysis (PCA)-based sensor fault reconstruction method was used for comparison under five fault conditions covering both single and simultaneous faults. Five variable scenarios were considered to validate the effectiveness of MLR-BI on HVACs with the limited variable information. Results indicated that the calibration accuracy of MLR-BI is over 99% under four conditions of the simulated case 1, which is about 6% and 8% higher than PCA and EC-BI respectively. For all the three variable scenarios of the simulated case 1, the calibration accuracy of MLR-BI is 99.65% on average. Especially in the four-variable scenario with limited variable information, MLR-BI shows the average calibration accuracy of 99.75% while PCA obtains 79.46% and EC-BI fails to work because of variable limitation. For the fault condition of the limitted-variable practical case 2, MLR-BI still outperforms the other two and obtains 97.1% calibration accuracy in two practical scenarios.

Imaging doses for different CBCT protocols on the Halcyon 3.0 linear accelerator – TLD measurements in an anthropomorphic phantom

IntroductionImage guided radiotherapy allows for particularly conformal tumour irradiation through precise patient positioning. Becoming the standard for radiotherapy, this increases imaging doses to the patient. The Halcyon 3.0 linear accelerator (Varian Medical Systems, Palo Alto, CA) requires daily imaging due to its geometry. For this reason, the accelerator is equipped with on-line kV and MV imaging. However, daily CBCT images required for irradiation apply additional radiation, which increases the dose to normal tissue and therefore can affect the patient's secondary cancer risk. In this study, actual organ doses were measured for the kV system, and a comparison of normal tissue doses for all available kV CBCT protocols was presented to demonstrate differences in imaging doses across entities and protocols. In addition, effective dose and secondary cancer risk from imaging are evaluated. Material and methodsMeasurements were performed with thermoluminescent dosimeters in an anthropomorphic phantom positioned according to each entity (brain, head and neck, breast, lung, pelvis). CBCT images were obtained, using all available pre-set protocols without further adjustment of the parameters. Measured doses for each position and each protocol were then compared and secondary cancer risk of relevant and specifically radiosensitive organs was calculated. ResultsIt was found that imaging doses for protocols such as Pelvis and Head could be reduced by up to half using the corresponding Fast and Low Dose modes, respectively. On the other hand, larger field sizes or the Large mode yielded higher doses than their initial protocols. Image Gently was found to spare normal tissue best, however it is not suitable for certain entities due to low image quality or insufficient projection data. DiscussionBy using appropriate kV-CBCT protocols, it is possible to reduce imaging doses to a significant extent and therefore spare healthy tissue. Combined with studies of image quality, the results of this study could lead to adjustments in workflow regarding the choice of protocols used in daily routine. This could prevent unnecessary radiation exposure and reduce secondary cancer risk.

Comparative study of machine learning methods and GR2M model for monthly runoff prediction

Monthly runoff time-series estimation is imperative information for water resources planning and development projects. This article aims to comparatively investigate the applicability of machine learning (ML) methods (i.e., Random Forest (RF), M5 model tree (M5), Support Vector Regression with polynomial kernel function (SVR-poly), and Support Vector Regression with the radial kernel function (SVR-rbf)) and the GR2M model for simulating the monthly runoff hydrograph. The models experimented at six runoff stations in Thailand’s Southern basin. Four performance criteria, including Nash-Sutcliffe Efficiency (NSE), Correlation Coefficient (r), Overall Index (OI), and Combined Index (CI), were utilized for model performance comparison. The finding results revealed that in stations with a low correlation coefficient (r) between input and output data sets, ML algorithms showed superior performance to GR2M. In particular, SVR-rbf showed outstanding performance over other methods. It expressed that SVR-rbf could manage the problem of low-quality data and simulate monthly runoff under limited available data.

Monte Carlo Tools for Assessing See Data Quality for Different Types of Analyses

After 43 years of Single-Event Effects (SEE) heavy-ion testing, there exists a large body of archival SEE results, almost all of which are intended for estimating on-orbit SEE rates. However, SEE data vary dramatically over a broad range of data-quality metrics. While variable SEE data quality may not be a major source of errors for SEE rates, other uses of the data—e.g., parametric estimation—are less forgiving of low-quality data. We apply a Monte Carlo (MC) tool using a simplified Generalized Linear Model (GLM) SEE data-fitting routine to assess how data quality affects SEE rate bounding, estimation and determination of fit parameters for SEE cross section (σ) vs. LET. The MC tool can generate 1000 MC events in about 1 minute on a standard laptop, making it suitable for on-the-fly testing decisions as well as evaluation of data quality.

VisQualdex: a Comprehensive Guide to Good Data Visualization

The rapid influx of low-quality data visualizations is one of the main challenges in today’s commu- nication. Misleading, unreadable, or confusing visualizations spread misinformation. Furthermore, they fail to deliver their message to the viewer. The scale of the problem is big enough that there already exist public fora gathering tens of thousands of users criticizing graphics and charts (reddit.com/r/dataisugly) made with obvious mistakes. Current attempts at data visualization appear mostly as simple and overgeneralized checklists, and often lack systematicity and versatility. The lack of proper tooling for evaluating data visualization quality further heightens the problem. Therefore, this paper proposes VisQualdex, a systematic set of guidelines for static data visual- ization. The codex categorization is based and inspired by the theory of Grammar of Graphics. It contains dozens of criteria designed to catch various errors and mistakes of different categories and magnitude. Furthermore, it has been peer-reviewed and tested by experts of data visualization, data science, graphics design, information technology and computer science. To apply theory in the real world and increase the practical impact of VisQualdex, this contribution also introduces a practical tool. The implementation of the guidelines is available in the form of the web server, https://visqual.onrender.com, developed as a single page application in JavaScript using Vue.js and Material Design principles.

Comparative Study of Machine-Learning-Based Methods for Log Prediction

Improving data quality during log preprocessing is an important task that can consume most of the time of the petrophysicist, with a high impact on the final interpretation. As part of the initiative to increase automation and homogeneity in the data completeness of logs in a field, we organized a systematic comparison of multiple regression models that provided successful predictions of wellbore logs. These approaches can be potentially valuable when extrapolating measurements available on a few wells to a more extensive set of wellbores, predicting low-quality data intervals, and increasing the availability of complete data sets. This study aims to compare the performance of three promising machine-learning (ML) methods when predicting one of the following curves: density, neutron porosity, and compressional slowness curves. We view the need to evaluate models that could provide answers even in the presence of multiple missing logs or logs with alteration, which is a common scenario in petrophysics. Because of that, we built a comparison based on three ML methods that can handle those issues: window-based convolutional neural network autoencoder (WAE), pointwise fully connected autoencoder (PAE), and eXtreme Gradient Boosting (XGBoost). We developed the PAE and WAE methods to handle challenging scenarios of interest, and we used the original implementation of XGBoost, which is already built to handle missing values. We compare the computational complexity, prediction errors [root mean square error (RMSE) and mean absolute error (MAE)], Pearson’s correlation, peak signal-to-noise ratio (PSNR), and the visual analysis of both high- and low-scale feature reconstruction, conducting the comparison in two field data sets. We also used the same methods to predict photoelectric factors and interpreted formation properties such as total organic content in multiple field data sets.

Deep Learning Methodology for Obtaining Ultraclean Pure Shift Proton Nuclear Magnetic Resonance Spectra

Nuclear magnetic resonance (NMR) is one of the most powerful analytical techniques. In order to obtain high-quality NMR spectra, a real-time Zangger-Sterk (ZS) pulse sequence is employed to collect low-quality pure shift NMR data with high efficiency. Then, a neural network named AC-ResNet and a loss function named SM-CDMANE are developed to train a network model. The model with excellent abilities of suppressing noise, reducing line widths, discerning peaks, and removing artifacts is utilized to process the acquired NMR data. The processed spectra with noise and artifact suppression and small line widths are ultraclean and high-resolution. Peaks overlapped heavily can be resolved. Weak peaks, even hidden in the noise, can be discerned from noise. Artifacts, even as high as spectral peaks, can be removed completely while not suppressing peaks. Eliminating perfectly noise and artifacts and smoothing baseline make spectra ultraclean. The proposed methodology would greatly promote various NMR applications.

The use of artificial intelligence in health care. Problems of identification of patients' conditions in the processes of detailing the diagnosis

The problems of using artificial intelligence in health care were discussed. The aim of the study. Assess the possibilities of using artificial intelligence in medicine right now. Most studies comparing the performance of AI and clinicians are not valid because the tests are not large enough or come from different sources. This difficulty could be overcome in the era of an open healthcare system. Indeed, open data and open methods are sure to attract a lot of attention as new research methods. It also highlights the idea that AI technologies can improve accuracy by incorporating additional data for self-updating, but automatically incorporating low-quality data can lead to inconsistent or inferior algorithm performance. The conclusion made is that the introduction of artificial intelligence into clinical practice is a promising field of development that is rapidly developing along with other modern fields of precision medicine. One of the fundamental issues remains the solution of ethical and financial issues related to the introduction of artificial intelligence

The Impact of Missing Data on Heart Rate Variability Features: A Comparative Study of Interpolation Methods for Ambulatory Health Monitoring

ObjectivesHeart rate variability (HRV) is a valuable indicator of both physiological and psychological states. However, the accuracy of HRV measurements taken by wearable devices can be compromised by errors during transmission and acquisition. These errors can significantly affect HRV features and are not acceptable for precise HRV analysis used for medical diagnosis. This study aims to address this issue by investigating the effectiveness of four different interpolation methods (Nearest Neighbour - NN, Linear, Shape-preserving piecewise cubic Hermite - Pchip, and cubic spline) in tackling missing RR values in real-time HRV analysis. Materials and MethodsIn this study, HRV signals were obtained from Electrocardiograms (ECG) through automatic detection and manually corrected by a specialist, resulting in high-quality signals with no missing or ectopic peaks. To simulate low-quality data acquisition, values were iteratively deleted from each HRV analysis window. The deleted values were then replaced using four different interpolation methods. Time and frequency domain features were computed from both the original and reconstructed signals, and the Mean Absolute Percentage Error (MAPE) was used to compare these features. ResultsResults showed that as the percentage of missing values increased, some interpolation methods were more suitable for RR time-series with a greater number of missing data. Furthermore, the study suggests that the impact of interpolation on HRV features varied across different features and that SDNN is the least affected by interpolation. In the time domain, nearest neighbour interpolation gives the best results for up to 50% missing data. Beyond this threshold, it seems better not to use any interpolation for RMSSD. In the frequency domain however, the lowest errors of HRV feature estimation are obtained using linear or Pchip interpolation. To achieve maximum performance, it is recommended to adapt the interpolation method to both the percentage of missing values and the targeted HRV feature. ConclusionResults highlight the importance of choosing the appropriate interpolation method to accurately estimate HRV features in real-time analysis. Overall, the Pchip interpolation seems to yield the best results on most HRV features as it preserves the linear trend of the data while adding very light waves. The findings can be beneficial in the development of more precise and reliable wearable devices for real-time HRV monitoring.

The role of data imbalance bias in the prediction of protein stability change upon mutation.

There is a controversy over what causes the low robustness of some programs for predicting protein stability change upon mutation. Some researchers suggested that low-quality data and insufficiently informative features are the primary reasons, while others attributed the problem largely to a bias caused by data imbalance as there are more destabilizing mutations than stabilizing ones. In this study, a simple approach was developed to construct a balanced dataset that was then conjugated with a leave-one-protein-out approach to illustrate that the bias may not be the primary reason for poor performance. A balanced dataset with some seemly good conventional n-fold CV results should not be used as a proof that a model for predicting protein stability change upon mutations is robust. Thus, some of the existing algorithms need to be re-examined before any practical applications. Also, more emphasis should be put on obtaining high quality and quantity of data and features in future research.

Effect of IMPACT team approach implementation processes on health supply chain management data use in Singida region: A quantitative study

The Government of Tanzania (GoT), in collaboration with development partners, has made significant improvements in the health supply chain since 2013 through developing of an electronic Logistics Management Information System (e-LMIS) and strengthening of the Logistics Management Services (LMS) component. Nevertheless, the use of reported data in the e-LMIS remained very low. In efforts to respond to low data use at all supply chain levels, the GoT introduced the Information Mobilized for Performance Analysis and Continuous Transformation (IMPACT) approach to reinforce data use culture for decision-making to improve the supply chain. Therefore, this study aimed to determine the effect of IMPACT team approach implementation processes on health supply chain data use in Singida region. This was a pre- and post-cross-section study whereby the quantitative data were obtained from e-LMIS stock imbalance reports. This data was processed using Microsoft Excel power query and analysed using Microsoft Excel 2019 pivot tables to get indicator results on availability, not-in-demand and overstocked health commodities. The overall pre (2018) and post (2020) intervention results for both Iramba district council (DC) and Singida municipal council (MC) were 66% and 81% for the availability of essential health commodities, 17.2% and 21% for not-in-demand health commodities and 17% and 20.7% for overstocked health commodities respectively. Iramba DC demonstrated a significant statistical increase in both indicators compared to Singida MC, as the P-value was less than 0.05 at a 95% confidence level (df=35). Conclusively, the IMPACT team approach positively influenced data use and availability of health commodities. However, it was not the case for overstocked and not-in-demand indicator results. Apart from the possible low quality of logistics data, further studies should explore other contributing factors behind the increasing trends of overstocks and not-in-demand health commodities in the region.