ฝึก Google Data Analytics

The data analysis process starts with Ask: defining the problem and questions before collecting or analyzing data. Prepare, Process, and Act come later.

Prepare involves gathering the right data and keeping it secure, organized, and ready for analysis. Dashboards, models, and deletion are not the main goals of this phase.

Analyze is where data is explored, cleaned, and transformed so patterns and conclusions can be identified. Question definition happens in Ask and reporting happens in Share/Act.

Primary data is collected firsthand for a specific analysis. Data from other organizations is secondary, and a spreadsheet or website can contain either type.

Structured data is organized in a fixed format such as rows and columns. Unstructured data, like text and images, has no fixed structure.

Customer reviews are text without a fixed row-and-column structure. Tables, databases, and CSV files are structured data.

Bias in data can come from sampling, collection, or labeling and can make conclusions less accurate. It is not a formula, duplicate, or encryption issue.

Data ethics means using data in ways that are fair, transparent, and respect privacy. It does not require publishing everything or collecting/storing data without limits.

Share is about presenting findings clearly so stakeholders can act. Collecting, cleaning, and selecting databases belong to earlier phases.

A cell is where a row and column meet and can hold a value or formula. Tabs, charts, and tables are different spreadsheet objects.

Pivot tables aggregate and rearrange data so patterns can be seen quickly. They do not delete rows, encrypt files, or only make charts.

XLOOKUP searches a range for a key and returns a matching value. SUM, COUNT, and TODAY perform different calculations.

TRIM removes leading, trailing, and repeated spaces. Deleting rows, rounding, and uppercasing are separate operations.

Conditional formatting changes cell appearance when values meet conditions, such as highlighting scores above a threshold. It is not row formatting, merging, or freezing.

COUNTIF counts cells meeting one condition. SUM adds values, MAX finds the largest, and COUNTA counts nonempty cells without a condition.

TRIM removes spaces, CLEAN removes nonprintable characters, and UPPER changes text to uppercase. NOW returns the current date and time.

Data validation controls allowed input, such as a list or number range. It is not spell check, formula summary, or a chart.

SPLIT separates text into multiple cells using a delimiter. Merging, splitting tabs, and removing duplicates are different operations.

A named range gives a range a memorable name for formulas. It is not a chart title, formula name, or sheet protection.

WHERE filters rows before grouping or sorting. ORDER BY sorts, GROUP BY aggregates, and LIMIT limits the number of rows.

SELECT * returns every column. SELECT ALL is not a standard wildcard, SHOW COLUMNS describes columns, and RETURN is not a SQL clause.

GROUP BY groups rows so aggregate functions such as COUNT, SUM, and AVG can be applied per group. Sorting, deduplication, and joining are separate operations.

COUNT returns the number of rows or non-null values. SUM adds numeric values, AVG averages them, and LENGTH measures text.

SELECT DISTINCT removes duplicate values from the returned rows. ORDER BY sorts, filtering handles conditions, and CREATE TABLE creates tables.

INNER JOIN keeps only matching rows from both tables. LEFT and RIGHT JOIN keep unmatched rows from one side, and FULL OUTER JOIN keeps both sides.

ORDER BY sorts rows by one or more columns. WHERE filters, GROUP BY groups, and LIMIT limits rows.

A primary key uniquely identifies each row and cannot be NULL. Foreign keys reference other tables, and indexes speed up searches.

LIMIT controls how many rows the query returns. Column selection uses SELECT, and LIMIT does not stop or resize tables.

CASE evaluates conditions and returns different values based on them. It is not automatic column creation, a join, or a count.

R is a programming language designed for statistics, data analysis, and visualization. It is not a spreadsheet, database, or storage service.

A data frame stores tabular data with named columns and rows. Vectors are one-dimensional, and plots are graphical output.

dplyr is part of the tidyverse and must be installed and loaded. It is not included in base R by default.

filter() keeps rows matching conditions. select() chooses columns, arrange() sorts rows, and rename() changes column names.

select() picks columns. filter() picks rows, mutate() creates or changes columns, and summarize() aggregates data.

mutate() adds new columns or modifies existing ones. filter(), arrange(), and joins perform different operations.

summarize() reduces data to summary values such as mean or count. It does not add rows, change types, or plot.

ggplot2 builds plots using a grammar of graphics. Connecting to databases, text cleaning, and scheduling are handled by other tools.

A vector holds one-dimensional values of the same type. Data frames are two-dimensional, plots are output, and queries are SQL.

read.csv() reads CSV data into R. write.csv() saves data, plot() creates graphics, and library() loads packages.

Tableau Desktop is a visualization tool for exploring and presenting data. It is not a database, formula, or programming language.

A data source defines the connection, fields, and joins for analysis. Dashboards, calculated fields, and worksheets are built on top of it.

A worksheet is where you build a single visualization. Dashboards combine worksheets, and reports are shared output.

Dimensions are categorical fields such as region or product that group data. Measures are numeric fields that are aggregated.

Measures are numeric fields that can be summed, averaged, or counted. Dimensions group and label data.

A calculated field uses a formula to create new values from existing fields. Imported fields, saved filters, and table calculations are different.

Filters limit which rows or marks appear in a view. They do not add columns, change chart types, or export files.

A dashboard combines multiple worksheets into one interactive view. It is not a single worksheet, source, or field.

A story presents dashboards and worksheets in order to guide the audience through findings. It is not a filter, chart, or table.

Show Me recommends visualizations from the fields you select. It does not hide sheets, create sources, or publish.

A dashboard presents important metrics and insights in one view. It does not store raw data, replace databases, or run queries by itself.

A KPI is a measurable value that shows progress toward a goal. It is not a password, index, or color scheme.

A line chart clearly shows changes over time. Pie charts compare parts of a whole, scatter plots show relationships, and treemaps show hierarchy.

Pie charts and stacked bars show composition. Line charts show trends, histograms show distributions, and box plots summarize spread.

Good dashboards focus on the metrics that matter to the audience. Too many metrics make insights harder to find.

Clear labels, appropriate charts, and a focus on key metrics improve readability. Decorative color overload can distract from the data.

Data storytelling weaves data and visuals into a clear narrative for stakeholders. It is not reading aloud, random charts, or hiding results.

Interactive filters let users adjust views by selecting values. They are not static images, queries, or saved reports.

Tooltips display additional context on hover. Titles, filters, and data sources are separate dashboard elements.

A useful report explains findings with evidence and next steps. Raw tables alone or unsupported opinions do not help stakeholders act.

Spreadsheet functions clean text, SQL queries relational data, Tableau builds interactive charts, and R scripts create reusable analyses.