Business Intelligence Exam 1 with complete solutions

Business Intelligence - Answer- the applications, infrastructure, and tools and practices that enable access and analysis of information to help make better decisions and improve and optimizw performance 3Vs - Answer- velocity, variety, and volume 4Vs - Answer- + veracity 6Vs - Answer- +Variability and value Big Data Criticisms - Answer- biases inherent to data theory doesn't matter gives rise to spurious relations Domain Knowledge Important because... - Answer- of the questions that should be asked, not bc of the answers A good research question is... - Answer- relevant to the situation at hand (answer will improve decision making and drive performance) can be answered with the data (feasibility) specific enough Components of a research question... - Answer- dependent and independent variables dependent variable - Answer- the subject, outcome, target, predicted variable independent variable - Answer- the feature, predictor variable subject - Answer- variable of interest predictor - Answer- the relevant info that can be used to learn about the variable of interest 3 types of analytics - Answer- descriptive predictive prescriptive descriptive analytics - Answer- focuses on what is or has happened, simply displaying the data we have at hand. ie. reports, dashboards, and scorecards questions: what is my retention rate? how much are we spending on advertisements? How has X affected Y over the past 4 years? Predictive Analytics - Answer- What's likely to happen? Making inferences, predictions, and forecasts with the data ie. regressions, classifications applications: fraud detection and bankruptcy forecast questions: is this email span? (questions we don't have the answers to but can make reasonable estimates based on data) Prescriptive Analytics - Answer- What should happen? (Try to engineer outcome based on descriptive and predictive analysis) Put findings of previous tasks to good use and think about how decisions impact everything else Use descriptive, predictive, and mathematical modeling to dynamically optimize certain outcome Tasks - Answer- What you do to find answers to the questions Types of tasks - Answer- regression classification clustering Regression - Answer- Estimate a numerical value of some variable Classification - Answer- categorize an observation (Yes/No) (Sunny/ Rainy/ Cloudy) Think in terms of bins results can be placed in Clustering - Answer- Grouping a set of observations/ individuals based on their similarity to one another ie. market segmentations, movie genres Types of problem approaches - Answer- supervised unsupervised supervised - Answer- those for which the target outcome is known for each observation data is labeled the model is trained by learning from examples Supervised Approaches - Answer- classification and regression Unsupervised - Answer- those in which the target outcome is not known for each observation data isn't labeled goal is to uncover patterns that weren't previously identified Unsupervised Approach - Answer- Clustering and grouping Types of Categorical data - Answer- nominal and ordinal Types of Numerical Data - Answer- discrete and continuous Categorical - Answer- data that can be sorted into groups/ categories ie. zipcode, phone number, area code Numerical - Answer- data that can be measured, sorted ie. income, height, travel time the average can be calculated Nominal - Answer- count but no order, just a named category (Gender, Eye Color) Ordinal - Answer- count and can be sorted, (Likert Scales, income brackets, and grades) Discrete - Answer- Data that can only take distinct and separate values (number of children, die rolls, age) Continuous - Answer- data can take on any value in the set of real numbers (height, age in decimal years, net profit) Characteristics of Descriptive Data - Answer- - methods of describing the dataset -allow you to make sense of the data -make conclusions about the data in order to make rational decisions -graphs are used to make sense of quantifications descriptive statistics - Answer- -measures of centrality -measures of dispersion -measures of shape -outliers Measures of centrality - Answer- finding the center of the data using mean, median, mode Mean - Answer- works well when there's no outliers Median - Answer- less prone to outliers so used with asymmetric data Measures of dispersion - Answer- Measures how much spread there is in the data from the measure of position metrics: range, variance, standard deviation Range - Answer- The simplest measure of variability, but can be misleading if the data is skewed and doesn't make full use of the available data Standard DEviation - Answer- a measure of variability of datapoint from their own mean. Lower SDs mean tighter dispersion around the mean 1 SD - Answer- 68% under the curve 2 SD - Answer- 95% under the curve 3 SD - Answer- 99.7% under curve Measures of shape - Answer- histograms (frequency distributions) Symmetric vs Asymmetric Unimodal vs Mutlimodal Skwedness help determine which descriptive stats are better use Outlier - Answer- A data point that is distinctly separate from the rest of the data. Purpose of Visualizations - Answer- Explanatory Analysis: gather relevant info and make sense of it and Presentation of Relevant Info : communicate your findings/ analysis Dashboard - Answer- a cognitive reporting tool that improves your span of control over a lot of data Scorecard - Answer- a type of dashboard that's more tactical and strategical Dashboard Design - Answer- clearly communicates key info and makes supporting info easily accessible Importance of Dashboards - Answer- h