Transcription
Moving Beyond the Spreadsheetfor AnalyticsFern HalperTDWI Research Director for Advanced AnalyticsAugust 14, 2014
Sponsor
SpeakersFern HalperResearch Director,Advanced Analytics,TDWIMatt MaddenSenior ProductMarketing Manager,Alteryx3
Agenda The analytics continuumLimitations of the spreadsheetTrends in analyticsThe analytics life cycleGetting started with analytics
The Analytics SpectrumVisualizationAdvancedAnalyticsOther BIDashboardsand reportsExcelOften becomes more algorithmic5
Benefits of spreadsheet Easy to use – for simple operations Nice charts Cheap6
Limitations of the spreadsheet Error proneData limitationsAnalysis limitationsManagement limitationsproscons7
The Analytics SpectrumVisualizationOther BIDashboardsand reportsExcelA lot of value to be gained!8AdvancedAnalytics
Analytics Drive Value9
Including advanced analtyics!45% MeasuredImpact10
Analytics trends worth noting1. Ease of use2. The emergence of the business analyst asmodel builder3. Data blending4. Analytic platforms11
1. Ease of Use Drag and drop interfaces- no scripting Iterative analysis More advanced analytics supported12
2. The emergence of the business analystTraditional Predictive User:Moving towards: Statistician/Modeler Critical thinker withknowledge of the businesse.g. business analysts
The Business Analyst Rules(source: TDWI Best Practices Report Predictive Analytics for Business Advantage, 2014)14
With an Evolving Skill SetThis ranks low15
3. Data BlendingCombining data from multiple sources withoutintegrating it into a data warehouse or othersystem of record16
Predictive analytics processProblemIdentification Framing problem Identifying data elementsModel ManagementData Access Evaluation/Monitoring Actual Management Integration BlendingDeploymentData preparation Sharing Scoring Operationalizing Cleansing TransformationData analysis Exploration Collaboration Validation17
Framing the problem Ability to formulate a question Comfortable creatively thinking in numbersand attributes18
Blending data togetherID NumberCust ID19
Data preparation Data quality Data transformation Data restructuring Data filtering20
Data analysis Many different kinds of tools– Reporting– Visualization– More advanced21
Analytics deployment and monitoring Visual workflow Schedulers22
4. Platforms are becoming popular83% in next 3 years23
Getting started Pick a problem with a known metric Start with a POC Work in steps24
Data Blending &Advanced AnalyticsMatt Madden- Sr. Product Marketing Manager, Alteryx 2014 Alteryx, Inc. Confidential
Analytics Solution Must DeliverBusiness Insight Through Those Who Know the BusinessRise of Analysts in Line of Business Understands business requirements Analytic thinker Accesses data, blendsand analyzesExpansion of Data Bigger datasets & more sources More Responsibility, ad-hoc analysis& reportingRelying on SpreadsheetsData Analyst Right Skillset – do more with less Not built for heavyprocessing/analysis 2014 Alteryx, Inc. Confidential26
Deeper Business InsightIntuitiveworkflow fordata blendingand advancedanalytics leadingto deeperinsights in hours,not weeksIn Hours Not WeeksThrough Intuitive Workflow 2014 Alteryx, Inc. Confidential27
The Alteryx Solution For Analyst EnablementAll Relevant DataPackaged Market &Customer DataEnrich 2014 Alteryx, Inc. ConfidentialAccess, cleanse, andblend data with uniquepackaged data optionsReportAnalyzeBlendTextRapid design ofpredictive analyticswith unique spatialunderstandingConsumerize the useof sophisticatedanalytics28
Data Preparation/Cleansing 60-80% analysts time Typically need IT, Data Specialist, orSQL to ensure the quality of data Restructure, reformat, re-condition Missing data or erroneousdata/manual entry Full dataset vs. Sample Preparing data for further analysis Reporting, database,visualization Traditional analyst tools limitedcapabilities 2014 Alteryx, Inc. Confidential29
Data Preparation/Cleansing in Alteryx Drag and drop icons to manipulateand alter formats of the data Transformations to help summarizeand aggregate your data Deal with missing data Incorporate spatial aspects in data Create formulas for if/thenstatements Filter on records that pertain toyour analysis Automatically set the field typesyou need And much more 2014 Alteryx, Inc. Confidential30
Joining Data– Most approaches Multiple spreadsheets Vlookups, formulas or manualprocesses Require writing SQL Traditional IT focused DataIntegration tools Slow processes Other tools may do simple joins thatfocus on one field Good, but what happens if youhave 3-4 data sources Joining spatial data may require evenanother tool 2014 Alteryx, Inc. Confidential31
Joining Data in Alteryx Drag and Drop Tools Join data of any type and anylevel Record and Row level Multiple key fields Spatial matching- based onspatial characteristicscustomer points, find nearest Non-identical data Fuzzy matching (match twodatasets based on nonmatching data-names, address,etc.) 2014 Alteryx, Inc. Confidential32
Advanced AnalyticsUsefulness of Spreadsheets Data Entry, Reporting Basic analysisWhere organizations/analystsstruggle Sales Forecasting Geo-spatial analysis Predictive AnalyticsAdvanced Analytics makes datamore manageable 2014 Alteryx, Inc. Confidential60 built-in tools for spatial and Rbased predictive analytics
DemoData Blending, Processingand Data Cleansing 2014 Alteryx, Inc. Confidential34
Major US Healthcare Provider uses Alteryx to pull data fromvarious sources, integrate, clean, and deploy reportsMedicare administratorsDeeper InsightsNow able to improveRegulatoryCompliance IssuesThrough data theypreviously were not ableto accessHours vs. WeeksEntire report now runs in37 Minutesvs.2 daysFor one section 2014 Alteryx, Inc. ConfidentialIntuitive WorkflowPreparing and blendingclient data from12 data files3.5M records35
Try it for yourself.Download the Alteryx Trial today to begincreating your own analysiswww.alteryx.com/download 2013 Alteryx, Inc. Confidential.3636
Questions?37
Contact InformationIf you have further questions or comments:Fern Halper, TDWIfhalper@tdwi.orgMatthew Madden, Alteryxmmadden@alteryx.com38
3 Speakers Fern Halper Research Director, Advanced Analytics, T
