Role & Scope
Contents
The U.S. Federal statistical system is a decentralized, interconnected network of 16 Recognized Statistical Agencies and Units, 24 Statistical Officials (across 24 major cabinet agencies), approximately 100 additional Federal statistical programs engaged in statistical activities, and several cross system interagency and advisory bodies.
The vision of the Federal statistical system is to operate as a seamless system, as stewards of much of the nation's most sensitive data, enabling greater evidence building, civic engagement, and public and private sector decision making.
Purpose: this resource is intended for users of Federal System System who need to:
- understand the structure of the Federal Statistical System (agencies, operations and geostatisitcal programs)
- determine what data are available to best meet application needs
- determine alternative sources of similar data to best meet application needs
- options available to access the data
- integrate these data with other data
- use the data; interpret and evaluate the data
- multi-sourced statistical release dates across selected agencies -- Calendar
Defining Federal Statistical Data
- something more than just data
- OMB OIRA authorization to collect data (Paperwork Reduction Act)
- authorized by Congress (a legal act)
- as described in the Federal Register
- for a defined geography (individual person or housing unit, political area, statistical area)
- collected/developed by a Federal agency
Derived Federal Statistical Data
- developed by a Federal agency
- developed using solely Federal statistical data (e.g. Congressional Communities)
Attributes of statistical data
- summary versus micro
Microdata refers to the raw, unaggregated data collected on individual entities, like a person or a household. In contrast, summary statistics are the results of calculations performed on that microdata to provide a concise overview of a dataset.
Microdata
Microdata is the most granular level of data, with each record representing a single unit of observation. Think of it as a spreadsheet where each row is a person and the columns contain their individual characteristics (age, income, education, etc.). Microdata is what a statistician would use to perform their own analysis.
| • | Key Characteristics: |
| o | Unit-level: Provides information on individual people, households, businesses, or other entities. |
| o | High Detail: Allows for complex, multi-variable analysis and the ability to explore relationships between different variables that might be lost in a summary. |
| o | Privacy Concerns: Due to its detailed nature, microdata must be carefully anonymized to protect the privacy of the individuals it represents. |
| • | Example: A census dataset containing a row for each person, with columns for their age, marital status, and employment. |
Summary Statistics
Summary statistics (also known as descriptive statistics) are values calculated from a dataset to describe its main features. They condense large amounts of information into a few key numbers or a simple table, making the data easier to understand at a glance.
| • | Key Characteristics: |
| o | Aggregated: The data is compiled into a single value or a small group of values. |
| o | Loss of Detail: You cannot use summary statistics to analyze individual-level relationships, as the original data has been boiled down. |
| o | Easy to Use: They are quick to calculate and provide a snapshot of the data's central tendency (mean, median), dispersion (standard deviation, range), and distribution (skewness). |
| • | Example: Calculating the average age or the median household income from a census dataset. |
- administrative versus statistical
Administrative data and statistical data differ primarily in their purpose and collection methods. Administrative data is collected for day-to-day operations, while statistical data is specifically gathered for analysis and research to draw conclusions about a larger population. Think of it this way: administrative data is the raw material, and statistical data is the refined product.
Administrative Data
Administrative data is information collected by government agencies, businesses, or other organizations as part of their routine operations. It is not originally intended for statistical analysis, but it can be used for that purpose. The main focus is on the specific identity of the individual or entity.
| • | Purpose: To manage, track, and administer programs and services. Examples include: |
| o | Tax records (for assessing and collecting taxes). |
| o | School enrollment data (for managing student attendance and resources). |
| o | Hospital patient records (for billing and treatment). |
| o | Voter registration lists (for conducting elections). |
| • | Strengths: Administrative data is often comprehensive, cost-effective to obtain, and can provide a complete count of a population or group since it's collected for everyone who interacts with the system. It can also provide a historical insight. |
| • | Weaknesses: Data quality may vary, and the information collected may not align perfectly with what's needed for statistical purposes. Changes in administrative procedures can affect data consistency over time. Additionally, the data can be restricted and may not include all the relevant background information. |
Statistical Data
Statistical data, on the other hand, is data collected with the explicit goal of producing official statistics. It is designed to be used for analysis to produce a summary of a group rather than to track an individual.
| • | Purpose: To understand trends, patterns, and relationships within a population or sample. It's used to inform policy decisions, evaluate programs, and conduct research. Examples include: |
| o | Census data (to provide a snapshot of a country's population). |
| o | Results from a survey on consumer spending habits. |
| o | Employment statistics collected by a national labor department. |
| o | Public health data on disease prevalence. |
| • | Strengths: Statistical data is typically collected using rigorous sampling methods and a well-defined conceptual framework, ensuring it is relevant, consistent, and representative of the target population. It is designed to be accurate and unbiased for the purposes of drawing inferences. |
| • | Weaknesses: Collecting statistical data can be very expensive and time-consuming. It may also suffer from non-response or sampling errors. |
- political versus statistical
Political and statistical areas are both ways to divide a country or region, but they differ in their purpose and criteria. Political areas are used for governance and administration, while statistical areas are designed for data collection and analysis.
Political Areas
Political areas are formal geographic divisions with specific legal and administrative functions. They are the building blocks of government and are used for purposes like elections, law enforcement, and providing public services. Their boundaries are defined by legal statutes and can be influenced by historical and political factors.
| • | Purpose: To govern and administer. |
| • | Examples: States, provinces, counties, cities, towns, and electoral districts. |
| • | Characteristics: |
| o | Fixed Boundaries: Defined by law and don't change frequently. |
| o | Legal Authority: They have their own governing bodies with legal jurisdiction and powers (e.g., city councils, county governments). |
| o | Historical and Political Influence: Boundaries can be shaped by historical events, natural features, or political agreements, which may not always align with current population or economic patterns. |
| o | Variation: A "county" in one part of the country may be vastly different in size, population, and function from a "county" in another. |
Statistical Areas
Statistical areas, on the other hand, are geographic regions defined by statistical agencies (like the U.S. Census Bureau) for the sole purpose of collecting, analyzing, and publishing data. Their boundaries are based on objective, data-driven criteria to create more meaningful comparisons.
| • | Purpose: To measure, compare, and analyze demographic and economic trends. |
| • | Examples: Metropolitan Statistical Areas (MSAs) and Micropolitan Statistical Areas (MISAs). |
| • | Characteristics: |
| o | Functional Delineation: They are defined by a core population center and the surrounding communities that have a high degree of social and economic integration with it, often measured by commuting patterns. |
| o | Fluid Boundaries: They are regularly reviewed and redefined to reflect current population and commuting trends. |
| o | No Legal Authority: They are simply for statistical purposes; they do not have their own governments or administrative functions. |
| o | Consistency: They provide a standardized way to compare urban and rural areas across a country, regardless of state or local political boundaries. For example, comparing the economies of the "New York-Newark-Jersey City MSA" and the "Los Angeles-Long Beach-Anaheim MSA" provides a more accurate "apples-to-apples" comparison than comparing the cities of New York and Los Angeles alone. |
- one point in time versus time series
The fundamental difference between "point in time" and "time series" data lies in their scope and purpose.
| • | Point-in-Time Data: This is a snapshot. It represents the state or value of something at one specific, discrete moment. Think of it as a single frame from a movie. It tells you "what" the data looked like at a particular instant, but it doesn't show you how it got there or where it's going. |
| o | Examples: |
| ▪ | The exact stock price of a company at 10:30 AM on a Tuesday. |
| ▪ | The total number of employees by county on December 31st of a given year. |
| ▪ | A report showing the number of homeless people in a city on a single night. |
| o | Key Characteristics: |
| ▪ | Static: It's a single data point or a set of data points for a specific moment. |
| ▪ | Non-additive: You can't add two point-in-time values together to get a meaningful sum (e.g., adding the number of employees on Monday to the number of employees on Tuesday doesn't make sense). |
| ▪ | Used for: Reporting on a specific condition, auditing, or comparing snapshots at different moments in time. |
| • | Time Series Data: This is a sequence of data points, ordered chronologically. It's the entire movie, not just one frame. The "time" element is a crucial variable, and the data is analyzed to understand how a variable changes over time. |
| o | Examples: |
| ▪ | The daily closing price of a stock over the past year. |
| ▪ | The sales figures for a retail store over a five-year period. |
| ▪ | Hourly temperature readings from a weather station. |
| o | Key Characteristics: |
| ▪ | Dynamic: It shows change and evolution over a period. |
| ▪ | Temporal ordering: The sequence of the data points is essential for analysis. |
| ▪ | Used for: |
| • | Trend analysis: Identifying long-term patterns, like an increasing or decreasing trend. |
| • | Seasonality: Spotting recurring patterns, such as sales spiking every December. |
| • | Forecasting: Predicting future values based on past behavior. |
| • | Identifying dependencies: Understanding how data points are related to previous data points. |
In summary, a point-in-time report answers the question, "What did the data look like at this specific moment?" A time series report, on the other hand, answers, "How has the data changed over time, and what patterns can we identify from that change?"
But what about ACS data
- survey versus model-based
"Survey data" and "model-based data" represent two different ways of obtaining information, each with its own strengths and weaknesses. The core distinction lies in how the data is generated.
Survey Data (Direct Data Collection)
Survey data is primary data collected directly from a group of people, or a "sample," to gather their opinions, behaviors, or characteristics. It is a direct measure of what a specific group thinks or does.
| • | How it's created: By asking people questions through methods like questionnaires, online surveys, phone interviews, or in-person interviews. |
| • | Strengths: |
| o | Direct and specific: It provides a direct view into the attitudes, beliefs, or actions of the people surveyed. |
| o | High fidelity: It is a raw, un-inferred representation of the respondents' answers. |
| o | Can capture nuance: Qualitative survey data (from open-ended questions) can provide rich, descriptive insights that are difficult to get from other data sources. |
| • | Weaknesses: |
| o | Sampling bias: The results are only as good as the sample. If the sample isn't representative of the larger population, the findings may not be generalizable. |
| o | Response bias: Respondents may not answer truthfully, may misunderstand questions, or may only be a certain type of person who chooses to take the survey. |
| o | Cost and time: Designing, distributing, and analyzing a comprehensive survey can be expensive and time-consuming. |
| o | Limited scope: It's a snapshot in time. It doesn't tell you about people who didn't respond or about the larger population outside of your sample. |
Model-Based Data (Inferred Data)
Model-based data, also known as modeled data, is secondary data that is not directly collected. Instead, it is created by using statistical or machine learning models to infer or predict information about a larger population based on a smaller, direct dataset (often survey data).
| • | How it's created: A data model is built using a known set of data (like survey responses) and then applied to a much larger dataset of people or entities (e.g., consumer behavior data, census data) to generate a probability or a score. The model essentially says, "Based on the people who answered the survey, here is how likely someone else with similar characteristics is to have the same opinion." |
| • | Strengths: |
| o | Scalability: A model can take a small amount of survey data and apply its insights to millions of individuals, making it highly scalable. |
| o | Predictive power: Models can be used for forecasting, targeting, and understanding the likelihood of certain behaviors. |
| o | Fills in gaps: It can provide insights for populations that are difficult or expensive to survey directly. |
| • | Weaknesses: |
| o | Assumption-based: The accuracy of the model-based data depends on the quality of the underlying model and the assumptions made during its creation. |
| o | Potential for error: If the model is flawed or the underlying data is biased, the resulting data will be inaccurate. |
| o | Less transparent: It's not a direct observation, but rather an inference. It's often difficult to fully explain why a model made a specific prediction. |
Analogy:
Imagine you want to know how many people in a city will vote in the next election.
| • | Survey Data: You call 500 randomly selected people in the city and ask them if they plan to vote. The data you get is the direct percentage of people in your sample who said "yes." |
| • | Model-Based Data: You take the survey data from your 500 respondents and create a model. You find that people who said "yes" tend to be over 40, own a home, and have voted in the last two elections. You then apply this model to a much larger database of all registered voters in the city. The model assigns a "propensity to vote" score to every person, predicting how likely they are to vote based on their characteristics. |
In this example, the survey data is the raw, factual starting point, while the model-based data is the scalable, predictive insight derived from it.
- summary statistic versus data about individual
The primary difference is that summary statistics condense a dataset into a few key numbers, while individual data provides the raw, unaggregated details for each element within that dataset.
Feature |
Summary Statistic |
Individual Data |
What it is |
A single value that describes a key characteristic of an entire group of data. |
The specific value for a single person, item, or observation. |
Purpose |
To provide a quick, high-level overview and make it easy to compare groups or understand the general trends. |
To retain all the original information and allow for detailed, case-by-case analysis. |
Examples |
Mean, median, mode, standard deviation, and range. |
The height of a single person, the exact score on one student's test, or a single sensor reading. |
Strength |
Efficient for large datasets, simplifies complex information, and is useful for reporting and communication. |
Captures outliers, preserves nuance, and is necessary for more complex modeling and analysis. |
Weakness |
Loses granular detail and can hide important variations or outliers within the dataset. |
Can be overwhelming and difficult to interpret on its own, especially with large datasets. |
- current/historical versus projection
The terms current, historical, and projection refer to different temporal states of data. The core difference is whether the data describes the past, the present, or a predicted future.
Current/Historical Data
This is actual data that has already occurred. It is a factual record of what has happened. The main distinction between current and historical data is a matter of recency.
| • | Current Data: Refers to information that is happening right now or is very recent. It's often referred to as "real-time data." This data is valuable for immediate, short-term decision-making. For example, a stock price at this exact moment or the current temperature. |
| • | Historical Data: Refers to data from the past, whether it's from minutes ago or decades ago. Historical data is critical for trend analysis, pattern recognition, and understanding how variables have behaved over time. For example, a company's sales figures from the past five years or the daily high temperature for every day last year. |
Current and historical data are the foundation for any kind of analysis because they are based on observable facts, not assumptions.
Projection
A projection is an estimate or prediction of a future trend or event. It is not factual data but rather an informed guess about what might happen, often based on an analysis of historical and current data. Projections are inherently speculative and are used for strategic planning and forecasting.
| • | How it's created: Projections are made by taking a set of assumptions and applying them to existing data. They often involve a "what-if" scenario. For example, a company might project its revenue for the next quarter by assuming a 10% increase in sales and using that assumption to calculate a future number. |
| • | Key Use Case: They are essential for financial planning, budgeting, and scenario analysis. For example, a company might create three different projections for the next year's revenue: a best-case scenario, a worst-case scenario, and a most-likely scenario. . |
- census versus survey
The fundamental difference between a census and a survey lies in the scope of data collection.
| • | A census aims for a complete enumeration of an entire population. It seeks to collect data from every single member of the group being studied. |
| • | A survey collects data from a representative sample of a population. It uses this smaller subset to make inferences and draw conclusions about the larger population. |
Breakdown of the key differences:
Feature |
Census |
Survey |
Scope |
The entire population. |
A sample of the population. |
Objective |
To get a comprehensive, detailed, and accurate count of the entire population. |
To get quick, cost-effective insights that can be generalized to the larger population. |
Accuracy |
High. Data is collected from everyone, so there is no sampling error. However, a census can still have some non-response error. |
Varies. The accuracy depends on how well the sample represents the population. There is always a margin of error. |
Cost & Time |
Very expensive and time-consuming. |
Much less expensive and faster. |
Frequency |
Typically conducted infrequently (e.g., every 10 years for a national census). |
Can be conducted frequently, even continuously (e.g., the U.S. Census Bureau's American Community Survey). |
Use Cases |
Government Planning: Allocating federal funding, determining legislative representation, and planning for infrastructure. Business: A full count of all employees and products for inventory purposes. |
Market Research: Understanding consumer preferences, satisfaction, or brand awareness. <br> Academic Research: Studying a specific behavior or opinion within a defined group. <br> Political Polling: Gauging public opinion on a particular issue. |
The U.S. Census Bureau: A Real-World Example
The U.S. Census Bureau provides a great example of how these two methods are used in tandem.
| • | The Decennial Census: This is a true census. Every 10 years, the Bureau attempts to count every single resident in the United States, as mandated by the Constitution. The goal is to get a complete picture of the population for political and economic purposes. |
| • | The American Community Survey (ACS): This is a continuous survey. The Census Bureau randomly samples a small percentage of households each year to collect more detailed information on a wide range of topics (e.g., income, education, housing). The data from this survey is used to provide timely statistics throughout the decade, without the time and expense of a full census. |
- special tabulation based on a census or ACS
A census is the primary, large-scale process of collecting data from every individual or entity within a specific population. A special tabulation is a customized data report created from that existing census data to meet a specific user's needs.
Think of it like this:
| • | The Census is the act of baking a giant, all-purpose cake. It's a massive, resource-intensive project that produces the foundational data—the "ingredients" and "batter" for everything else. The results are the standard, broad-based data products released to the public. |
| • | A Special Tabulation is a slice of that cake, cut and decorated to a customer's specific order. It's not a new cake; it's a unique report generated by re-analyzing the raw, underlying census data. For example, while the census might release standard tables showing the number of households by income and by age, a special tabulation could provide a custom table that cross-references households by both income and the age of the householder for a specific, non-standard geographic area. |
In essence, a census is the source of the data, while a special tabulation is a customized product created from that source, often for a fee, to answer a very specific question that the standard public data releases cannot address.
- secure versus non-secure data
The concepts of "secure data" and "non-secure data" are central to cybersecurity and data privacy. The core difference lies in the protective measures applied to the data.
Non-Secure Data is data that lacks adequate protection. It is vulnerable to unauthorized access, theft, or corruption. Think of it as a letter sent through the mail in an unsealed envelope. Anyone handling it can easily read the contents.
Characteristics of Non-Secure Data:
| • | Plaintext or Unencrypted: It is stored or transmitted in a format that is easily readable by anyone who gains access to it. |
| • | Poor Access Controls: There are no or weak restrictions on who can view, modify, or delete the data. This might include weak passwords, no user authentication, or giving everyone full access. |
| • | Vulnerable Storage and Transmission: It is stored on systems or transmitted over networks that are not protected by firewalls, antivirus software, or encryption. |
Secure Data is data that is protected from unauthorized access, use, or disclosure. It is an active state of being protected, not just a label. Using the analogy above, it's like a letter sealed in an envelope, with a locked box for delivery, and a signature required for the recipient.
Characteristics of Secure Data:
| • | Encryption: The data is scrambled into a coded format (ciphertext) that is unreadable without a decryption key. This is a primary method of securing data both "at rest" (when it's stored on a server) and "in transit" (when it's being sent over a network). |
| • | Strong Access Controls: Access is granted only to authorized individuals based on their roles and a "need-to-know" basis. This involves strong passwords, multi-factor authentication, and robust identity management. |
| • | Data Integrity: Measures are in place to ensure the data cannot be modified or corrupted without being detected. |
| • | Confidentiality: Policies and technologies are used to ensure the data is only accessible to authorized people, thereby protecting its privacy. |
The Interplay with Data Privacy
It's crucial to understand that data security is a component of data privacy, but they are not the same.
| • | Data Security is about the how: the technical and organizational safeguards to protect data from threats. |
| • | Data Privacy is about the what and why: the rights of individuals to control their personal data, and the rules and regulations (like GDPR and CCPA) that govern how that data is collected, used, and shared. |
A system can be technically very secure (e.g., using strong encryption), but still violate data privacy if it collects and uses data in ways the user did not consent to. Conversely, you can't have data privacy without data security because a data breach would expose all private information.
- determining reliability of the data
Determining the reliability of data involves assessing its quality, consistency, and trustworthiness. It's about ensuring the data is fit for its intended purpose and that you can make sound decisions based on it. Reliable data is a foundational component of good data quality and a prerequisite for valid conclusions.
Here are the key factors to consider when determining data reliability:
1. Accuracy and Validity
| • | Accuracy: Does the data reflect the real-world truth? Incorrect data, such as a typo in a name or an outdated address, is not accurate and therefore not reliable. You can test accuracy by comparing your data to a known, trustworthy source. |
| • | Validity: Does the data adhere to predefined rules and formats? For example, is a date field formatted as a date, or is a zip code field only populated with numbers? Data is invalid if it violates the rules of the system it's in. While not the same as reliability, invalid data is a sign of unreliability. |
2. Consistency and Uniqueness
| • | Consistency: Is the data uniform across different sources and over time? If a customer's name is "John Smith" in one system and "Jon Smyth" in another, the data is inconsistent. Unreliable data often arises from data silos and lack of standardization. |
| • | Uniqueness: Are there duplicate records? Duplicates can skew analysis by over-representing certain data points. Ensuring data uniqueness is a crucial step in maintaining reliability. |
3. Completeness and Timeliness
| • | Completeness: Is all the necessary information present? Missing values, like a blank phone number field in a customer record, can make the data unreliable for certain analyses. |
| • | Timeliness: Is the data up-to-date and current? Outdated data, such as a sales report from a year ago for a current marketing campaign, is not reliable for making timely decisions. |
4. Source and Lineage
| • | Source: Where did the data come from? Data from a trusted, official source (e.g., a government agency) is generally more reliable than data from an unverified or anonymous source. |
| • | Lineage: How has the data been processed and transformed? You should be able to trace a data point back to its original source, noting all changes made along the way. A clear data lineage helps you trust the data's integrity. |
5. Methodology and Governance
| • | Methodology: How was the data collected? Flaws in the collection method, such as a biased survey question or a broken sensor, can lead to unreliable data. You must understand the methodology to judge the data's quality. |
| • | Governance: Are there clear rules and responsibilities for data management? A strong data governance framework, which includes policies for data entry, storage, and security, is essential for ensuring ongoing data reliability. Without it, human error and lack of accountability can quickly compromise data quality. |
- NAICS codes -- data by type of business
The North American Industry Classification System (NAICS) provides a standardized way to classify businesses by their primary economic activity. It's a hierarchical system that allows for data to be organized and analyzed at different levels of specificity, from broad economic sectors to very specific industries.
How NAICS Codes Work
The NAICS system uses a six-digit code to classify businesses. Each digit adds more detail:
| • | First two digits: Represent the economic sector, such as "Manufacturing" (31-33) or "Retail Trade" (44-45). |
| • | Third digit: Defines the subsector. For example, within "Manufacturing," a 311 would indicate "Food Manufacturing." |
| • | Fourth and fifth digits: Further break down the industry into more specific industry groups and industries. |
| • | Sixth digit: Designates a specific national industry, allowing for classification that is tailored to the economies of the U.S., Canada, and Mexico. |
Data by Type of Business
The primary purpose of NAICS is to enable federal agencies in the U.S., Canada, and Mexico to collect and publish statistical data about the economy. This allows for a variety of data, such as employment, wages, and economic output, to be compiled and analyzed by business type.
For example, a government agency or a market research firm can use NAICS codes to:
| • | Determine the number of businesses within a specific industry. |
| • | Track economic trends over time, such as which sectors are growing or shrinking. |
| • | Compare the economic performance of different industries or regions. |
| • | Identify potential competitors or customers within a specific business sector. |
The NAICS system replaced the older Standard Industrial Classification (SIC) system to better reflect the modern service-based economy and to provide a more consistent basis for data comparison across the three North American countries.
- disclosure/suppression coded into data
"Suppression" coded into data refers to the intentional removal or alteration of data points to protect an individual's identity or to prevent the misinterpretation of statistically unreliable information. Instead of the actual value, a specific code is entered into the dataset to indicate that the original data has been suppressed.
Why Data is Suppressed
Data suppression is a critical practice in data publishing, particularly by government agencies and research institutions. The two primary reasons for it are:
1. Confidentiality and Privacy 🔒
This is the most common reason for data suppression. It is used to protect personally identifiable information (PII). When data is granular or specific to a small group, it becomes possible to identify individuals by cross-referencing different data points. For example:
| • | Small Group Sizes: If a report shows the average income for a rural town's single-family households, and there are only two such households, the data could be used to figure out the income of those specific families. To prevent this, the data is suppressed. |
| • | Extreme Values: A single, unusually high or low value in a small group could reveal a person's information. For instance, if a public health report shows that a certain rare disease has one case in a specific, small town, that single data point could identify the individual. |
To avoid these privacy breaches, government agencies often apply minimum cell size rules, suppressing data for any group below a certain threshold (e.g., fewer than 5 or 10 individuals).
2. Statistical Reliability 📈
Data is also suppressed when the numbers are too small to be statistically reliable. A small sample size can lead to a result with a large margin of error, making it unreliable for drawing conclusions. For example, if a survey of 10 people in a county finds that 8 of them voted for a certain candidate, it's not a reliable indicator of the county's voting behavior. To prevent this from being misinterpreted, the data would be suppressed. This ensures that only robust and meaningful statistics are published
How Suppression is Coded
Instead of leaving a blank space, which could be misinterpreted as a zero, a specific code or symbol is used to indicate that the data has been intentionally suppressed. Common codes include:
| • | "D" for "Disclosurue" (used by the U.S. Census Bureau for economic data). |
| • | "S" for "Suppressed." |
| • | An asterisk (*) or a hyphen (-), often with an accompanying footnote explaining the reason for the suppression. |
| • | A zero (0), which can be confusing but is sometimes used when the exact count is very low to prevent a data user from being able to determine the exact value of the original, suppressed data. |
By using these codes, data publishers can provide useful information while protecting individual privacy and maintaining the integrity of their data. For a user, understanding these codes is crucial to avoid drawing incorrect conclusions from the data.
- jam values
In statistics and data reporting, especially with data from the U.S. Census Bureau's American Community Survey (ACS), "jam values" are special codes used in place of actual numeric data. They represent a situation where the exact value is not available because it either falls below a certain threshold or is an open-ended number.
These codes are used for two primary reasons:
| 1. | To protect confidentiality: When a data value represents a very small number of individuals or households, releasing the exact number could compromise their privacy. A jam value is used to prevent this disclosure. |
| 2. | To indicate an open-ended category: Sometimes, a data category is not a precise number but a range. A jam value like "$2,000+" is used to show that the value is at or above a certain point. Another example is "10-" which indicates a value of 10 or fewer. |
By using these codes, data publishers can provide useful information while protecting individual privacy and maintaining the integrity of their data. They are essentially a form of data suppression that is specifically formatted to represent these unique situations.
Focus on Application Programming Interface (API)
API stands for Application Programming Interface and provides a developer with programmatic access to a proprietary software application.
An API is software that makes it possible for application programs to interact with each other and share data.
An Application Programming Interface (API) is a set of rules and protocols that allows different software applications to communicate with each other. It acts as an intermediary, enabling one application to access the functionality or data of another in a standardized way.
Think of an API as a waiter in a restaurant. You (the customer) have a request (an order for food). The kitchen (the server) has what you want, but you can't go directly into the kitchen and get it yourself. The waiter (the API) takes your request, delivers it to the kitchen, and then brings the finished product back to you. The waiter is a defined interface; you know exactly how to interact with them (placing your order), and you don't need to know how the kitchen works.
Key Concepts
| • | Request and Response: The communication between applications is a simple request-and-response cycle. One application sends a request to an API, and the API sends a response back, which can be data, a status message, or an action confirmation. |
| • | Endpoints: An API's functionality is organized into endpoints, which are specific URLs that represent different resources or actions. For example, a weather API might have an endpoint for get_current_weather and another for get_forecast. |
| • | Data Formats: APIs typically exchange data in standard formats like JSON (JavaScript Object Notation) or XML (Extensible Markup Language), which are easy for both computers and humans to read. |
How APIs are Used
APIs are the backbone of modern software development and are used in a vast range of applications:
| • | Web Development: When you book a flight on a travel website, it uses APIs to retrieve real-time data from various airlines. When you see a map on a website, it's often powered by an API from a service like Google Maps. |
| • | Mobile Apps: Mobile apps rely heavily on APIs to fetch data from remote servers. For example, a social media app uses an API to get your friend's posts and upload your photos. |
| • | Internet of Things (IoT): Smart devices like thermostats and speakers use APIs to send data to and receive commands from a central server. |
| • | Business Integration: APIs allow different business systems, like a customer relationship management (CRM) system and an accounting system, to share data seamlessly. |
See the scope of agencies covered in this document.