# # Google Analytics - 2 Ask - Module 2 - Make data-driven decisions ## 0. Overview >[! quote] In analytics, data drives decision-making, and this is your opportunity to explore data of all kinds and its impact on all sorts of business decisions. You’ll also learn how to effectively share your data through reports and dashboards. ### Learning Objectives >[! info] Module 2 Learning Objectives >- [ ] Discuss the use of data in the decision-making process >- [ ] Compare and contrast [[Data-driven decision-making]] and [[Data-inspired decision-making]] >- [ ] Explain the difference between [[Quantitative data]] and [[Qualitative data]] including reference to their use and specific examples. >- [ ] Discuss the importance and benefits of [[Dashboard]] and [[Report]] to the data analyst with reference to [[Tableau]] and [[Spreadsheet|spreadsheets]] >- [ ] Differentiate between [[data]] and [[Metric|metrics]], giving specific examples >- [ ] Demonstrate an understanding of what is involved in using a mathematical approach to analyze a problem. --- ### Glossary for Module 2 - [[Algorithm]] - [[Big data]] - [[Dashboard]] | [[Data-inspired decision-making]] - [[Metric]] | [[Metric goal]] - [[Pivot chart]] | [[Pivot table]] | [[Problem types]] - [[Qualitative data]] | [[Quantitative data]] - [[Report]] | [[Return on investment (ROI)]] | [[Revenue]] - [[Small data]] <hr> ## 1. Understand the power of data ### 1.1 Video _Data and Decisions_ (1:00) Quick intro video of what is to come. :) ### 1.2 Video: _How data empowers decisions_ (5:00) How does data play into the decision process. Case study: Search for restaurants near me: a decision made by data analysis [[Data-inspired decision-making]]: Explores different data sources to find out what they have in common. [[Algorithm]]: a process or set of rules to be followed for a task. 90% of the data that exists was created within the last few years. Just having data is not enough. ==Need to interpret data to turn it into information== [[context]] adds to the value of the information ==When we use and apply information, it becomes knowledge.== ### 1.3 Reading: _Data trials and triumphs_ review the difference and similarities between [[Data-driven decision-making]] and [[Data-inspired decision-making]] #### Data-driven decisions. - marks the use of data as the basis for the decision. - depends on the quality and quantity of available data - pitfalls when data is insufficient or is biased - can create an over relaiance on certain data types too. - historical data - ignorance to qualitative data as well - Example of [[Data-driven decision-making]] is A/B testing. #### Data-inspired decision making - often includes qualitative data - can avoid pitfalls of pure data-driven decisions - Example is CSAT + other qualitative factors inform the decision or strategy the Customer Service manager deploys. Another example was about how Pepsi changed their approach to data ... the case study was linked to a [Think with Google](https://www.thinkwithgoogle.com/marketing-strategies/data-and-measurement/pepsi-digital-transformation/) article. ==important to remember that you can have solid data and still make the wrong choices.== Another reason having skilled analysts can help businesses make sense of their data and make good choices. EG New coke launch was NOT successful despite having lots of solid data. EG. NASA's loss of Mars orbiter due to poor cross-functional communication. One program used newtons to calculate force while another program used pounds to calculate force. both programs were used to control the spacecraft resulting in RUD in the martian atmosphere at a cost of $125M ### 1.4 Video: _[[Qualitative data]] and [[Quantitative data]]_ (4:00) 2 kinds of data being discussed; [[Quantitative data]] and [[Qualitative data]] #### Quant the specific and objective measures of numerical factors - the what , the how many, the how often of a problem. #### Qual describes subjective or explanatory qualities of data. - useful in answering why questions Putting them together. Quantitative data can provide us with charts and graphs based on the measurements while Qualitative data will enable high level "why" helps to add context. EG product reviews is a good example. I wonder what kinds of analysis Proof bread is doing ?? ### 1.5 Reading: _[[Qualitative data]] and [[Quantitative data]] in business_ In the reading I learned from a hypothetical case study where I am in the role of an analyst at a movie theater chain. The key takeaways here is that context is important and a decision that impacts the business often uses both quantitative data and qualitative data that explains the why ### 1.6 Practice Quiz: _Test your knowledge on the power of data_ Untimed practice quiz. Basically going ver new terminology covered in the subsection. #### Results 100% <hr> ## 2. Follow the evidence ### 2.1 Video: _The big reveal: Sharing your findings_ (5:00) Here talking abou two types of data presentation tools [[Report]] and [[Dashboard]] Pro and con of each of them General diff is that report is static, dashboard is live. #### Reports Pros - great at giving periodic high level snapshots - easy to design and consume - reflect data that is clean and sorted Cons - Reports need continual maintnence - not live - not visually appealing #### Dashboards Pros - dynamic, automatic and interactive - more stakeholder access - low mx - visually appealing Cons - large labor cost to design and implement - can require significant mx overhead - can be confusing / TMI #### ReportDemo Spreadsheets with [[Pivot table]] >[! Definition] >A data summarization tool used to sort, recognize, group, count, total, or average data. #### Dashboard Demo Tableau demo There is still a lot to learn before getting into the specifics of building reports or dashboards. ### 2.2 Video: _Data vs metrics_ (3:00) Topic is specifically about the difference between [[data]] vs [[Metric]] >[! Definition] >A single, quantifiable type of data that is used for measurement example metric [[Return on investment (ROI)|ROI]] = net profit / cost of investment. net profit is another metric and cost of investment. Different industries use all kinds of different metrics. This is [[Metric goal]] which is set by the company. ### 2.3 Reading: _Tools for visualizing data_ The reading is an overview of tools that an analyst might use, their advantages and their disadvantages. #### Spreadsheets - go to for static charts/graphs on small data - can also be used to clean, sort, and filter a data set on the spreadsheet. - another strong aspect is that popular spreadsheet applications are tightly integrated with other useful productivity applications like word processors or presentation applications - The point of this is that you can link a chart to a presentation. The chart can update automatically when the source data is updated in the spreadsheet app. very handy if you are making the same presentation on the same data (but the data changes) all the time. - #### Tableau - used to create powerful and interactive visualizations. - because of this, this is a tool of choice for many analysts when building dashboards. - downside is that this tool has a learning curve as well. ### 2.4 Reading: _Design compelling dashboards_ Great link to see some real world [tableau](https://www.tableau.com/learn/articles/business-intelligence-dashboards-examples) examples understanding the usefulness of a dashboard works broadly for both [[stakeholder|stakeholders]] and for data analysts in four areas. - Centralization - Data analysts: Works on a single source of data - Stakeholders: enables a comprehensive view of data - Visualization - Data analysts: Show an update live incoming data in real time - Stakeholders: enables to identify changes, trends, and patterns more quickly. - Insightfulness - Data analysts: pulls relevant data from different datasets. - Stakeholders: Can understand the story behind the numbers, enables [[Data-driven decision-making]] - Customization - Can create custom views dedicated to a specific project, person, presentation ... - Stakeholders: Drill down to specialized area of concern. Important to remember that dashboards can update live data as long as the underlying data structure does NOT change. - Change in data structure is bad, very bad. - requires redesign. #### Data dashboard creation process >[! cue] [[Dashboard creation process]] There are four steps >[!cue] [Requirements Gathering Worksheet from Looker](https://s3.amazonaws.com/looker-elearning-resources/Requirements+Gathering+Worksheet.pdf) 1. Identify the stakeholders - Who needs to see the data and how will it be used. - Use the requirements worksheet to help! - ask effective questions. 2. Design the dashboard - Use clear header to label the information - Add short text descriptions to each visualization - Show most important information at the top - Document your metadata for others to be able to use your custom dataset. 3. Create mockups - sketching is good to test your ides. 4. Select visualizations - Selecting the visualization - Select the visualization that aligns with the story you want to tell. - Change in value over time, line or bar charts - How a part contributes to a whole, pie/donut chart. >[!cue] Tableau [Dashboard showcase](https://www.tableau.com/solutions/gallery) and [viz of the day](https://www.tableau.com/solutions/gallery) 5. Create filters as needed - use the [Filter Actions](https://help.tableau.com/current/pro/desktop/en-us/actions_filter.htm) page for inspiration ### 2.5 Self-reflection: _Go deeper into dashboards_ Three most common categories of dashboard are [[Strategic Dashboard]] : focuses on long term goals and strategies at the highest level of metrics, think of KPI's and overall strategies. [[Operational Dashboards]] short-term performance tracking and intermediate. These are the most common types of dashboards. [[Analytical Dashboards]] : consists of the datasets and the mathematics used in these stets ### 2.6 Practice Quiz: _Test your knowledge of following the evidence_ 4 questions untimed. #### Results 100% Need to remember that dashboards are low mx (maintenence. ) <hr> ## 3. Connect the data dots ### 3.1 Video: _Mathematical Thinking_ (4:00) Use a mathematical thinking approach. Look at a problem break down step by step to see the relationship patterns in data >[! quote] >This kind of thinking helps us figure out the best tools for analysis because it helps us figure out the different aspects of a problem and choose the best logical approach. Specific metrics over a short time-period = [[Small data]] DEMO using SQL There are a number of metrics that might show us the pattern that can give us insight. In my opinion this was a bit of toy example though. No explanation on what a mathematical thinking framework might look like. IF it was there, I missed it. ### 3.2 Reading: _[[Big data]] and [[Small data]]_ The reading discusses some key differences challenges and benefits between [[Big data]] and [[Small data]] #### Big data usaully is less specific, large, and covers a long period of time. is complex by nature usually is stored in a [[Database|database]] and is accessed by using a [[Query|query]] or queries Takes a lot of resources to collect, store, secure, manage, sort, and represent visually. in it's natural state, big data, tends to need to be broken down before analysis can occur. #### Small data describes very specific data made up of measurements/[[Metric]]. usually organized into [[Spreadsheet|spreadsheets]] simple to collect, store, manage, sort and represent. Usually is a manageable size already. #### Challenges and benefits working with big data Challenges: - too much data. There is a lot of noise and it could be hard to uncover the signals - because signals are harder to reach, decision making time fames slow down. - needed data isn't always accessible quickly. - Turnkey solutions can still struggle on big data, can lead to bias. - gaps in big data solutions exist at many organizations. Benefits: - large amounts of data, though noisy, often hide game-changing insights. - Great for trend analysis and CSAT #### Three V's (four v's ) of big data >[!cue] [[4 - V's of big data]] >[! info] The four V's of big data >- Volume: the amount of data >- Variety: the different kinds of data >- Velocity: how fast can the data be processed >- Veracity: the quality and reliability of the data ### 3.3 Practice Quiz: 4 questions, untimed. #### Results 100% <hr> ## 4. Module 2 challenge ### 4.1 Glossary - [[Algorithm]] - [[Big data]] - [[Dashboard]] | [[Data-inspired decision-making]] - [[Metric]] | [[Metric goal]] - [[Pivot chart]] | [[Pivot table]] | [[Problem types]] - [[Qualitative data]] | [[Quantitative data]] - [[Report]] | [[Return on investment (ROI)]] | [[Revenue]] - [[Small data]] ### 4.2 Quiz: _Module 2 challenge_ Timed test 40 minutes 80% to pass #### Results 5:20 test time. 100% ROI was a question. Not giving away the answer, but I was not expected to be tested on a very specific metric. It is a good metric and is applicable across industries and business types. Nearly universal. <hr> ## Summary and takeaways >[! summary-top] Module 2 Key Take Aways >1. [[Big data]] not always great, but consider [[4 - V's of big data]]. >2. [[Metric]] are important, begin to consider how something can be measured. >3. [[Quantitative data]] and [[Qualitative data]] are both important. >4. A [[Pivot table]] is a summarization tool that can sort, count, average, group and total [[Small data]]