Extracting Information from Data
Introduced| English | Meaning |
|---|---|
| information/ˌɪnfəˈmeɪʃn/ | Data or communicated content that conveys meaning or reduces uncertainty. |
| data sets/ˈdeɪtə sets/ | data sets |
| patterns/ˈpætnz/ | patterns |
| trends/trendz/ | The general direction in which a quantity or pattern changes over time. |
| Filtering/ˈfɪltərɪŋ/ | Filtering |
| Cleaning/ˈkliːnɪŋ/ | Cleaning |
| correlation/ˌkɒrɪˈleɪʃn/ | A statistical association between variables, which does not by itself establish causation. |
| Causation/kɔːˈseɪʃn/ | A relationship in which a change or event helps bring about another change or event. |
| Metadata/ˌmetəˈdeɪtə/ | Data describing other data, such as a file's author, creation time, format or dimensions. |
| visualization/ˌvɪʒuːəlaɪˈzeɪʃn/ | visualization |
| bias/ˈbaɪəs/ | A systematic tendency for a measurement, estimate or selection process to differ from the true or representative value. |
From raw data to information
- Raw data by itself is just numbers and text.
- The goal is to turn it into useful information 信息 — something that answers a real question.
- Programs can scan huge data sets 数据集 far faster than a person.
- By scanning, they find patterns 模式 (things that repeat) and trends 趋势 (changes over time).
Turning raw data into something that answers a real question produces:
Difficulty 1/5 · 100 coins + a small bonus on first completion today
Information is data made useful — patterns and trends answering a question.
Match each term to its meaning.
Difficulty 2/5 · 140 coins + a small bonus on first completion today
Programs scan data sets to find both.
Preparing messy data
- Real data is messy, so it must be prepared first.
- Filtering 筛选 keeps only the rows you care about and hides the rest.
- Cleaning 清洗 fixes problems — removing duplicates, filling missing values, correcting mistakes.
- Preparing data this way makes the later analysis meaningful and trustworthy.
Filter the rows you need, then count, sum or average them
Filtering or cleaning?
Filtering keeps only the rows you care about; cleaning fixes problems in the data (duplicates, blanks, mistakes) before analysis.
Removing duplicate rows and filling missing values is part of cleaning data.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
Cleaning fixes problems so the analysis can be trusted.
Correlation is not causation
- A correlation 相关 means two things change together (ice-cream sales and sunburns both rise in summer).
- Causation 因果 means one thing actually causes the other.
- A correlation does not prove causation.
Don't confuse the two. Ice cream does not cause sunburns — hot, sunny weather causes both. Two things rising together can share a hidden cause, so a correlation alone never proves that one causes the other.
Ice-cream sales and sunburns both rise in summer. This shows:
Difficulty 1/5 · 100 coins + a small bonus on first completion today
Hot weather causes both — a shared cause, not one causing the other.
"Data about data", like a photo's date and location, is called ______.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
Metadata helps organize, search, and locate information.
Surveying only people who downloaded an app about whether the app is popular introduces:
Difficulty 1/5 · 100 coins + a small bonus on first completion today
The collection method slants the result — a biased sample.
Metadata, visualization, and bias
- Metadata 元数据 is "data about data" — a photo's date, camera, and location. It helps organize and search.
- A visualization 可视化 turns data into a picture (a chart or graph) so a pattern is obvious at a glance.
- Always ask how the data was collected: bias 偏差 is a slant that pushes conclusions one way.
Biased survey. A company asks "Is our app popular?" but surveys only people who already downloaded it. Almost all say yes — but unhappy users never took the survey, so the happy result is misleading.
Programs turn raw data sets into information, finding patterns and trends. Prepare data by filtering (keep relevant rows) and cleaning (fix problems). Beware: a correlation never proves causation, and bias in how data was collected can mislead. Metadata and visualization help organize and reveal insight.