CPSC provides accessibility to recalls via a recall database. The information is publicly available to consumers and businesses as well as software and application developers.

CPSC's National Electronic Injury Surveillance System (NEISS) is a national probability sample of hospitals in the U.S. and its territories. Patient information is collected from each NEISS hospital for every emergency visit involving an injury associated with consumer products.

The PhysioNet/Computing in Cardiology Challenge 2020 uses data from several distinct databases. The sources primarily describe the characteristics and format of these datasets as they are provided to participants, rather than outlining explicit preprocessing steps performed by the challenge organisers. It is noted that participants may need to handle existing data issues.

Here is information regarding the characteristics of each dataset and the reasoning behind their names:

Data Characteristics and Preprocessing

The data for this Challenge is provided in WFDB format, consisting of binary MATLAB v4 files for the ECG signal data and text files in WFDB header format that describe the recording and patient attributes, including the diagnosis. A key aspect of the Challenge is that there are bound to be some errors or debatable labels in each database, and participants are expected to "work out how to deal with these issues". The quality of labels can vary, with some databases having human overread machine labels and others having single or multiple human labels.

Here are the specific details for each:

• CPSC Database and CPSC-Extra Database

  • Preprocessing/Characteristics: This training set comprises two subsets: 6,877 recordings (male: 3,699; female: 3,178) and 3,453 recordings (male: 1,843; female: 1,610). The ECG recordings last from 6 seconds to 60 seconds and were sampled at 500 Hz.
  • Name Origin: "CPSC" stands for the China Physiological Signal Challenge in 2018 (CPSC2018), from which this data was sourced. The "CPSC Database" is the public data from this challenge, while the "CPSC-Extra Database" refers to unused data from CPSC2018.

• St Petersburg INCART 12-lead Arrhythmia Database

  • Preprocessing/Characteristics: This public dataset consists of 74 annotated recordings, extracted from 32 Holter records. Each record is 30 minutes long and contains 12 standard leads, each sampled at 257 Hz.
  • Name Origin: The name explicitly states its origin as the "St Petersburg INCART 12-lead Arrhythmia Database", indicating it's an arrhythmia database from St Petersburg.

• PTB Diagnostic ECG Database and PTB-XL electrocardiography Database

  • Preprocessing/Characteristics: These are two public databases from the Physikalisch Technische Bundesanstalt (PTB).
    • PTB Diagnostic ECG Database: Contains 516 records (male: 377, female: 139). Each recording was sampled at 1000 Hz.
    • PTB-XL electrocardiography Database: Contains 21,837 clinical 12-lead ECGs (male: 11,379 and female: 10,458). These recordings are 10 seconds long with a sampling frequency of 500 Hz.
  • Name Origin: "PTB" is an acronym for Physikalisch Technische Bundesanstalt, the institution from which these databases originated. "PTB-XL" signifies a larger, perhaps "extra large," version of the PTB database.

• The Georgia 12-lead ECG Challenge (G12EC) Database

  • Preprocessing/Characteristics: This training set contains 10,344 12-lead ECGs (male: 5,551, female: 4,793). The recordings are 10 seconds long with a sampling frequency of 500 Hz.
  • Name Origin: The name directly indicates its geographical origin as a "Georgia database", which represents a "unique demographic of the Southeastern United States". "G12EC" is an abbreviation of "Georgia 12-lead ECG Challenge".

• Undisclosed American Database

  • Preprocessing/Characteristics: This source contains 10,000 ECGs, all of which are retained as test data and are not released for training. No specific sampling frequency or length is provided in the general description for this particular database, but it is implied to be similar to other ECG recordings.
  • Name Origin: The name "Undisclosed American database" clearly states that its specific origin or provider is not revealed and that it is from an American location, distinct geographically from the Georgia database. The test data is not released to the public to prevent overfitting and ensure fair competition.

The CPSC Clearinghouse Online Query Tool is a public database maintained by the U.S. Consumer Product Safety Commission. It provides access to incident data related to deaths, injuries, and hazards associated with consumer products. The data is compiled from multiple sources such as death certificates, news reports, medical examiners, healthcare providers, and direct consumer submissions. Users can search, filter, and download incident records covering the most recent ten years. The database supports research, analysis, and reporting on consumer product safety trends by providing detailed information such as product codes, hazard types, victim demographics, and incident narratives.

On March 11, 2011, the U.S. Consumer Product Safety Commission launched SaferProducts.gov. This site hosts the agency's new Publicly Available Consumer Product Safety Information Database. On SaferProducts.gov, consumers can submit reports of harm or reports of potential harm. After a short amount of time for review by the agency and named manufacturer, these reports go live on SaferProducts.gov and are searchable by the public. The public also can export search results. The Application Protocol Interface (API), to open the published SaferProducts.gov data to developers and businesses so that the information in SaferProducts.gov can be accessed by an even greater number of consumers online and on mobile devices.

1) Data Introduction • The U.S. Consumer Product Safety Recalls Dataset is a structured and refined version of official recall data from the U.S. Consumer Product Safety Commission (CPSC), including the name, date, type of risk, manufacturer, number of injuries, and URL of the product recalled for safety concerns.

2) Data Utilization (1) U.S. Consumer Product Safety Recalls Dataset has characteristics that: • This dataset reflects actual product safety incidents, and is a mixture of categorical data such as product name, manufacturer, risk type, and numerical/time series data such as number and date of injury. • Each row represents a single recall case and is optimized for various analysis tasks such as NLP, time series prediction, and risk classification. (2) U.S. Consumer Product Safety Recalls Dataset can be used to: • Risk Type Classification Model: It can be applied to the development of NLP models that automatically classify risk types (fire, food, etc.) by BERT/Random Forest, etc., using product name and explanatory text. • High-risk product pattern analysis: By analyzing monthly recall frequency trends, repeat recall status by manufacturer, and injury rate correlation, it can contribute to quality supervision policy establishment.

Traffic analytics, rankings, and competitive metrics for cpsc.gov as of September 2025

Beginning in July 2000, the National Center for Injury Prevention and Control (NCIPC), Centers for Disease Control and Prevention (CDC) in collaboration with the United States Consumer Product Safety Commission (CPSC) expanded the National Electronic Injury Surveillance System (NEISS) to collect data on all types and causes of injuries treated in a representative sample of United States hospitals with emergency departments (EDs). This system is called the NEISS-All Injury Program (NEISS-AIP). The NEISS-AIP is designed to provide national incidence estimates of all types and external causes of nonfatal injuries and poisonings treated in U.S. hospital EDs. Data on injury-related visits are being obtained from a national sample of U.S. NEISS hospitals, which were selected as a stratified probability sample of hospitals in the United States and its territories with a minimum of six beds and a 24-hour ED. The sample includes separate strata for very large, large, medium, and small hospitals, defined by the number of annual ED visits per hospital, and children's hospitals. The scope of reporting goes beyond routine reporting of injuries associated with consumer-related products in CPSC's jurisdiction to include all injuries and poisonings. The data can be used to (1) measure the magnitude and distribution of nonfatal injuries in the United States; (2) monitor unintentional and violence-related nonfatal injuries over time; (3) identify emerging injury problems; (4) identify specific cases for follow-up investigations of particular injury-related problems; and (5) set national priorities. A fundamental principle of this expansion effort is that preliminary surveillance data will be made available in a timely manner to a number of different federal agencies with unique and overlapping public health responsibilities and concerns. Also, annually, the final edited data will be released as public use data files for use by other public health professionals and researchers. NEISS-AIP data on nonfatal injuries were collected from January through December each year except the year 2000 when data were collected from July through December (ICPSR 3582). NEISS AIP is providing data on approximately over 500,000 cases annually. Data obtained on each case include age, race/ethnicity, gender, principal diagnosis, primary body part affected, consumer products involved, disposition at ED discharge (i.e., hospitalized, transferred, treated and released, observation, died), locale where the injury occurred, work-relatedness, and a narrative description of the injury circumstances. Also, major categories of external cause of injury (e.g., motor vehicle, falls, cut/pierce, poisoning, fire/burn) and of intent of injury (e.g., unintentional, assault, intentional self-harm, legal intervention) are being coded for each case in a manner consistent with the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) coding rules and guidelines. NEISS has been managed and operated by the United States Consumer Product Safety Commission since 1972 and is used by the Commission for identifying and monitoring consumer product-related injuries and for assessing risk to all United States residents. These product-related injury data are used for educating consumers about hazardous products and for identifying injury-related cases used in detailed studies of specific products and associated hazard patterns. These studies set the stage for developing both voluntary and mandatory safety standards. Since the early 1980s, CPSC has assisted other federal agencies by using NEISS to collect injury- related data of special interest to them. In 1990, an interagency agreement was established between NCIPC and CPSC to (1) collect NEISS data on nonfatal firearm-related injuries for the CDC Firearm Injury Surveillance Study; (2) publish NEISS data on a variety of injury-related topics, such as in-line skating, firearms, BB and pellet guns, bicycles, boat propellers, personal water craft, and playground injuries; and (3) to address common concerns. CPSC also uses NEISS to collect data on work-related injuries for the National Institute of Occupational Safety and Health (NIOSH), CDC. In 1997, the interagency agreement was modified to conduct the three-month NEISS All Injury Pilot Study at 21 NEISS hospitals (see Quinlan KP, Thompson MP, Annest JL

An up-to-date list of entities that have been accredited to assess conformity with product safety rules

When CPSC is involved in a civil or criminal investigations into violations of the Consumer Products Safety Act the Commission publishes final determinations and those penalties are recorded in the Civil and Criminal Penalties Database. You can search this database for records by civil or criminal penalties as well as by company, product, and fiscal year.

For more than 45 years, the CPSC has operated a statistically valid injury surveillance and follow-back system known as the National Electronic Injury Surveillance System (NEISS). The primary purpose of NEISS is to collect data on consumer product-related injuries occurring in the United States. CPSC uses these data to produce nationwide estimates of product-related injuries.NEISS is based on a nationally representative probability sample of hospitals in the U.S. and its territories. Each participating NEISS hospital reports patient information for every emergency department visit associated with a consumer product or a poisoning to a child younger than five years of age. The total number of product-related hospital emergency department visits nationwide can be estimated from the sample of cases reported in the NEISS.NEISS Data Highlights CPSC staff produces annual NEISS Data Highlights reports that summarize injury data by major product groups, product subgroups, age, and sex. These reports are published in mid to late April and are a good entry point to understanding the kind of data available through NEISS.Data highlights are divided into 4 folders - Overview, Top 20 Product Injury Estimates by Age, Top 20 Product Injury Estimates by Sex, and Top 20 Product Injury Estimates by Age and Sex. Each PDF represents 1 year, indicated in the title.Files downloaded May 23, 2025.

The Monthly Progress Report (MPR) is provided by recalling firms to report on the progress of the recall. The MPR reports recalled products at the Manufacturer, Distributor, Retailer, and Consumer level on a monthly basis.

CPSC's epidemiological data include reports of incidents involving death, injury, or potential injury that are associated with consumer products. The online Clearinghouse posts summary information from death certificates (DTHS), medical examiner reports (MECAP reports), reports published on Saferproducts.gov, Newsclips, and other submissions from consumers, healthcare professionals, state, federal, and local agencies (IPII), and public safety entities.

This dataset provides a structured and cleaned version of consumer product recall data from the official U.S. Consumer Product Safety Commission (CPSC).

Each row represents a product that has been recalled due to potential safety risks. The dataset includes:

1.Product Name 2.Recall Date 3.Hazard Type 4.Manufacturer 5.Number of Injuries (if reported) 6.Direct URL to the recall report

The dataset is ideal for:

1. Machine Learning Projects: Predict risk levels or classify product categories based on hazards
2. Trend Analysis: Identify monthly/seasonal spikes in recalls or high-risk categories
3. Natural Language Processing (NLP): Work with real-world titles and descriptions to train classification or summarization models
4. Public Health & Policy Research: Understand which industries and demographics are more affected by product failures

Project Ideas Here are some impactful project ideas based on your dataset:

1.Hazard Type Classification Task: Classify the hazard type based on the recall title using NLP Models: Logistic Regression, Random Forest, BERT

2.Monthly Trend Forecast Task: Forecast number of recalls or injuries per month Tools: Time-series models (ARIMA, Prophet), Rolling averages

3.High-Risk Manufacturer Detection Task: Identify manufacturers with frequent or severe recalls Insight: Can be used for accountability or supply chain audits

4.NLP-Based Risk Summarizer Task: Build an NLP model that generates short summaries of recall risk from the title Tools: BART, T5, GPT-style fine-tuning

5.Interactive Dashboard Tool: Tableau / Power BI / Plotly Dash / Streamlit

Views: Recalls over time, hazard types by manufacturer, map of affected regions (if extended)

Source Official Site: www.cpsc.gov/Recalls Data collected via web scraping and cleaned using Python + Pandas

Created by: Vaibhav Agrawal For educational and research use only

For all products regulated by the CPSC, the Commission issues a Letter of Advice (LOA) when there is a violation of a mandatory standard

CPSC 2018The first dataset is a preprocessed version of the CPSC 2018 dataset, which contains 6877 ECG recordings. We preprocessed the dataset by resampling the ECG signals to 250 Hz and equalizing the ECG signal length to 60 seconds, yielding a signal length of T=15,000 data points per recording.For the hyperparameter study, we employed a fixed train-valid-test split with ratio 60-20-20, while for the final evaluations, including the comparison with the state-of-the-art methods and ablation studies, we used a 10-fold cross-validation strategy.The raw CPSC 2018 dataset can be downloaded from the website of the PhysioNet/Computing in Cardiology Challenge 2020.(License: Creative Commons Attribution 4.0 International Public License).PTB-XL (Super-Diag.)The second dataset is a pre-processed version of PTB-XL, a large multi-label dataset of 21,799 clinical 12-lead ECG records of 10 seconds each. PTB-XL contains 71 ECG statements, categorized into 44 diagnostic, 19 form, and 12 rhythmic classes. In addition, the diagnostic category can be divided into 24 sub- and 5 coarse-grained super-classes. In our pre-processed version, we utilize the super-diagnostic labels for classification and the recommended train-valid-test splits, sampled at 100 Hz. We select only samples with at least one label in the super-diagnostic category,without applying any further preprocessing.The raw PTB-XL dataset can be downloaded from the PhysioNet/PTB-XL website.(License: Creative Commons Attribution 4.0 International Public License).

In 1992, the National Center for Injury Prevention and Control (NCIPC), a unit of the Centers for Disease Control and Prevention (CDC), established an interagency agreement with the U.S. Consumer Product Safety Commission (CPSC)to begin collecting data on nonfatal firearm-related injuries by using the National Electronic Injury Surveillance System (NEISS), the primary data system of CPSC. This ongoing special study is commonly called the "CDC Firearm Injury Surveillance Study". These data provide the basis for national estimates of nonfatal firearm-related injuries and nonfatal BB/pellet gun-related injuries treated in hospital emergency departments in the United States. Beginning in July 2000, NCIPC, in collaboration with CPSC, expanded NEISS to collect data on all types and causes of injuries treated in a representative sample of hospitals. This system is called the "NEISS All Injury Program (NEISS AIP)". These data provide the basis for national estimates of all types of nonfatal injuries treated in hospital emergency departments in the United States.

Users can access reports including data pertaining to product safety. Topics include but are not limited to: children’s products, electrocutions, fires, fireworks, sports and recreation, and submersion. Background The consumer product related statistic database in maintained by the US Consumer Product Safety Commission. The US Consumer Product Safety Commission produces reports on a variety of topics covering injuries and death. User functionality Data is presented in report or abstract form and can be downloaded in PDF formats by clicking on the publications link. All reports and abstracts use United States Data. Data Notes The data sources are cle arly referenced for each article. The most recent publications are from 2008. There is no indication on the site when the data will be updated.

Abstract (en): These data were collected using the National Electronic Injury Surveillance System (NEISS), the primary data system of the United States Consumer Product Safety Commission (CPSC). CPSC began operating NEISS in 1972 to monitor product-related injuries treated in United States hospital emergency departments (EDs). In June 1992, the National Center for Injury Prevention and Control (NCIPC), within the Centers for Disease Control and Prevention, established an interagency agreement with CPSC to begin collecting data on nonfatal firearm-related injuries to monitor the incidence and characteristics of persons with nonfatal firearm-related injuries treated in United States hospital EDs over time. This dataset represents all nonfatal firearm-related injuries (i.e., injuries associated with powder-charged guns) and all nonfatal BB and pellet gun-related injuries reported through NEISS from 1993 through 2000. The cases consist of initial ED visits for treatment of the injuries. Cases were reported even if the patients subsequently died. Secondary visits and transfers from other hospitals were excluded. Information is available on injury diagnosis, firearm type, use of drugs or alcohol, criminal incident, and locale of the incident. Demographic information includes age, sex, and race of the injured person. United States hospitals providing emergency services. Stratified probability sample of all United States hospitals that had at least six beds and provided 24-hour emergency services. There were four hospital size strata (defined as very large, large, medium, and small, based on the number of annual ED visits) and one children's hospital stratum. From 1993 through 1996, there were 91 NEISS hospital EDs in the sample. In 1997, the sampling frame was updated so that from 1997 through 1999, the sample included 101 NEISS hospital EDs. In 2000, one NEISS hospital dropped of the system so there were 100 NEISS hospital EDs in the sample. In 1997, CPSC collected firearm-related cases using the "old" and "new" NEISS hospital samples for a 9-month period. This dataset includes data from the "new" sample. The overlapping "old" sample is not included. Comparisons of weighted estimates based on the "old" and "new" samples indicated a difference of about 1 percent in the overall national estimate using these samples. The characteristics of firearm-related cases from these two overlapping samples were also very similar. 2005-11-04 On 2005-03-14 new files were added to one or more datasets. These files included additional setup files as well as one or more of the following: SAS program, SAS transport, SPSS portable, and Stata system files. The metadata record was revised 2005-11-04 to reflect these additions.2003-09-16 The 2000 data have been added to the cumulative data. The codebook and SAS and SPSS data definition statements have been updated to reflect these changes.2002-09-19 The 1999 data have been added to the cumulative data and a variable was removed. The codebook and data definition statements have been updated to reflect these changes.2001-05-18 The 1998 data have been added to this study, and the codebook has been updated to reflect these changes. Funding insitution(s): United States Department of Health and Human Services. Centers for Disease Control and Prevention. National Center for Injury Prevention and Control. United States Department of Justice. Office of Justice Programs. Bureau of Justice Statistics.

ObjectiveTo evaluate the National Electronic Injury Surveillance System’s (NEISS) comparability with a data source that uses ICD-9-CM coding. MethodsA sample of NEISS cases from a children’s hospital in 2008 was selected, and cases were linked with their original medical record. Medical records were reviewed and an ICD-9-CM code was assigned to each case. Cases in the NEISS sample that were non-injuries by ICD-9-CM standards were identified. A bridging matrix between the NEISS and ICD-9-CM injury coding systems, by type of injury classification, was proposed and evaluated. ResultsOf the 2,890 cases reviewed, 13.32% (n = 385) were non-injuries according to the ICD-9-CM diagnosis. Using the proposed matrix, the comparability of the NEISS with ICD-9-CM coding was favorable among injury cases (κ = 0.87, 95% CI: 0.85–0.88). The distribution of injury types among the entire sample was similar for the two systems, with percentage differences ≥1% for only open wounds or amputation, poisoning, and other or unspecified injury types. ConclusionsThere is potential for conducting comparable injury research using NEISS and ICD-9-CM data. Due to the inclusion of some non-injuries in the NEISS and some differences in type of injury definitions between NEISS and ICD-9-CM coding, best practice for studies using NEISS data obtained from the CPSC should include manual review of case narratives. Use of the standardized injury and injury type definitions presented in this study will facilitate more accurate comparisons in injury research.