click below
click below
Normal Size Small Size show me how
ACFM 329
Exam 2
| Term | Definition |
|---|---|
| How do you create a Data Analysis Project Plan (hint it is 5 steps) | 1. Focus on objective 2. Select data strategy 3. Select analysis strategy 4. Consider risks 5. Embed controls |
| Data analysis plan increases the likelihood that analysis results accurate described what 4 things? | What happened, diagnose why it happened, predict what is likely to happen, or prescribe best course forward |
| Data is considered appropriate for analysis when they are: | Relevant, available, and characteristics match the analysis method requirements |
| Generated within the organization & can be more easily controlled & verified by the organization | Internal Data |
| Obtained from sources outside of organization & are somewhat riskier to use since we often cannot know if data are accurate or complete, but can provide insights that internal data alone cannot provide | External Data |
| Collection of data columns & rows available for analysis | Data Set |
| Individual columns in a data set, one unique characteristic | Fields |
| If source of data was a database the columns are called __________ | Attributes |
| Rows in a data set which represent collection of columns that hold descriptions of a single occurrence of data sets purpose | Records |
| What are the 2 raw data fields? | Measured Data & Non-Measured Data |
| Data created or captured by controlled process capturing value of data & can be discrete or continuous data | Measured Data |
| Data created automatically by computer or company policy for control & is discrete data | Non-Measured Data |
| Data created when one or more fields in a particular record have any number of mathematical operators applied, & often are derived from using data in another field within same record (never across rows) & can be discrete or continuous data | Calculated Data Fields |
| Labeled or named data that can be sorted into groups according to specific characteristics | Categorical (Nominal) Data |
| Ordered or ranked categorical data | Ordinal Data |
| Ordinal data that have equal & constant differences between observations & an arbitrary 0 point | Interval Data |
| Interval data with a natural 0 point | Ratio Data |
| When a data risk involves non-representative sample selections, what is the control? | Verify representativeness |
| When a data risk involves outlier data points, what is the control? | Perform histogram/quartile analysis |
| When a data risk involves dirty data - variety of errors, what is the control? | Verify integrity of data set & clean up issues |
| Can chosen analysis strategy answer specific objective questions? Is measurement scale of data appropriate for selected analysis strategy? | What 2 questions should we consider when designing data analysis strategy? |
| Central tendency, frequency distributions, data dispersions, visualizations, correlation, calculations, & regression | What are some potential analysis strategies for descriptive & diagnostic analytics? |
| Trend lines, linear regression analysis, optimization models, & what-if analyses | What are some potential analysis strategies for prescriptive & predictive analytics? |
| Often involve interdisciplinary data & knowledge | Accounting Information Systems |
| What are the 4 types of analytics objectives that are commonly used? | Descriptive, diagnostic, predictive, & prescriptive |
| Select data that captures accounting information system's: | Performance, vulnerability, & errors |
| Focused on increasing ________________________ & improving _______________________ | Competitive advantage & Organization's operations |
| Use statistics to determine risks to account balances & unusual transactions | Auditing |
| Most often use categorical, interval, & ratio data All 4 types of analytics are used Use statistics to identify opportunities & problems with profitability, liquidity, & business valuation | Financial Accounting |
| Want to improve planning, operational control, & decision-making to support organization's mission & strategies | Managerial Accounting |
| Predictive & prescriptive | Tax Accounting |
| Prerequisite for effective data analytics, but many organizations lack a structured approach to detecting & resolving dirty data | Clean, reliable data |
| Human error, poor system design, lack of internal controls, time pressure, merging systems, failure to appreciate data quality, lack of accountability, & improper incentives | Causes of Dirty Data |
| Missing, invalid, duplicate, & incorrectly formatted data | Dirty Data |
| Focuses on what is happening | Descriptive Data |
| Involve many of the same analyses as descriptive, but focus is on explaining why | Diagnostic Data |
| Use historical data to create models that estimate a future value or outcome | Predictive Data |
| Help determine which option is likely to produce best outcome given objective | Prescriptive Data |
| There is a 10-step process to detect & resolve dirty data, what are the first 5 steps? | Understand business process Analyze source & processing What should be in the data? Scan sample of recent data Build descriptive reports for each table |
| There is a 10-step process to detect & resolve dirty data, what are the last 5 steps? | Develop & document expectations for data within each field Document expectations for relationships among fields Test each expectation by creating exception reports Evaluate exceptions & identify root causes Act to clean data & address root causes |
| Learn how data reflects actual operations, perform walkthrough, & understanding is the basis for knowing what data to expect & identifying data issues | Understand Business Process |
| Where does data originate?, what internal controls govern data input process?, & is data processed or editable later? | Analyze Source & Processing |
| What should tables/fields be? & lack of documentation may be a signal that dirty data issues are likely to be present | What Should Be in the Data? |
| Helps get a sense of what the actual data looks like & identify any potential issues before proceeding further in data cleansing process | Scan Sample of Recent Data |
| Determine missing data, identify min, avg, max for numerical/date fields, count frequency of categorical variables, & provides an overall analytical review to become familiar with the data & detect any potential dirty data issues | Build Descriptive Reports for Each Table |
| Unique identifiers, allowable values, value ranges, & can better identify & resolve any issues that do not meet those expectations, leading to a cleaner data set | Develop & Document Expectations for Data Within Each Field |
| Identify relationships between fields across tables, specify expected data dependencies, note expected patterns across fields, document data integrity rules, & can establish a comprehensive set of data quality criteria to assess integrity of overall data | Document Expectations for Relationships Among Fields |
| Identify expectation violations, exceptions reveal data issues, empty exception report indicates compliance, & organizations can effectively identify & surface areas of data that do not meet documented expectations, setting stage for further investigation | Test Each Expectation by Creating Exception Reports |
| Review exceptions, determine if exceptions reflect dirty data, identify root causes of dirty data, & organizations can take targeted actions to clean current data & prevent future dirty data issues | Evaluate Exceptions & Identify Root Causes |
| Clean current data, address root causes, consider patterns, people, & processes, establish accountability & make changes, & organizations can ensure data quality & enable better informed decision making going forward | Act to Clean Data & Address Root Causes |
| Process of generating additional knowledge from data that is relevant for analysis purposes | Information Modeling |
| What are the steps in the Information Modeling Process? | Data (input) Algorithm (transform) Information (output) |
| Sets of instructions that transform data into information, link between input & output; intrinsic to accountants | Algorithms |
| Output of additional knowledge gained from data | Information |
| What are the types of information fields? | Calculated Column & Measure |
| New data set column with values that are calculated based on values in other columns | Calculated Column |
| Aggregated data information field that can be used for analytical purposes | Measure |
| Performing data analysis requires determining numbers to analyze (measures) & how to analyze them (dimensions) | Structured Approach |
| Fields that can slice them | Dimensions |
| Define dimensions of ______, _________, & ___________ in transactions to help make accounting sense of data | Who, What, When |
| Preferred data structures for data analyses | Star Schemas |
| Data structure that shows relationships among different entities in data set | Data Model |
| What are the 2 goals of creating measures & dimensions? | Create rich set of measures for fact table Rich set of dimensions that can break down, or slice, measures in many ways |
| Dimension table that describes agents involved in the business transaction data | Who Table |
| Data relationship that links agents to specific business transactions | Participates Relationship |
| Dimension table that describes resources involved in a business transaction | What Table |
| Data relationship between resources & specific transactions | Flows Relationship |
| Dimension table that describes when the data from business transaction occurred | When Table |
| Data relationship that links calendar to specific business transaction | Occurs Relationship |
| Which patterns implement information modeling algorithms? | Within-Table Numeric Calculation Within-Table Text Calculation Within-Table Classification Across-Table Calculation Single-Column Aggregation Filtered Aggregation Measure Hierarchies |
| Creates a new field (calculated column) from one or more numeric columns, or fields in same table | Within-Table Numeric Calculation |
| Creates a new field (calculated column) from text fields in same table | Within-Table Text Calculation |
| Can be applied to a data set unlimitedly, data binning, & most rely on Boolean logic | Within-Table Classification |
| Numbers are grouped into smaller numbers of "bins" | Data Binning |
| New column is created from existing data, tables must be linked, questions to consider: are there properly defined relationships between tables, what is nature of relationship & what type of join should be considered, & what cardinalities apply? | Across-Table Calculation |
| Measure applies mathematical operation to all values of single column, "one measure, multiple analyses principle" | Single-Column Aggregation |
| One measure can be broken down into many different ways | One Measure, Multiple Analyses Principle |
| Used when new, more complex measure is created using existing measures. Especially useful when calculating ratios & benchmarks | Measure Hierarchies |
| Which patterns help develop & implement accounting information models? | How many Participates | Transaction-Who Flows | Transaction-What Occurs | Transaction-When Who-What-When Star Schema Integrated Star Schema |