Icon

Rossmann - Analiza szeregów czasowych

Data Preparation:Aggregate daily Rossmann sales data into monthly data Data Preparation -Automation of appending data from disparate filesandJoining weather data with sales data Data Preparation:Calculating new variables based on existing column in sales data Data Preparation:Filtering data and selecting 4 store for example analysis Data Preparation:Calculating new variables based on lagged values of weather (+1,+2) and sales and customer data (-12) Data Preparation:Appending together spllited store data and filtering unused in modellingrows Modelling metanodes 2015 prediction Segmentation:Basic segmentation of stores according to their features and aggregated transactional data 1. train1,017,209 rows9 colsstorechange date from default string storage get Year, Month etc from datevar:SalesPerCustomerdata auditcreate flow variablesvar:Cost (as 0.8*Sales+ random() )Select 10 storesexclude Open = 0and Sales = 0CompetitionSince ... DateCompet and PromoPromotionSince ... Datevar:Promo2SinceMonth(formula proxy)AggregateStores to Monthchange StateHolto integerreplace inString columnStateHolidaycreate Date (Monthly)change date from default string storage colored (use Vis Settings for adjustment)Storesand AssortmentColor: Assortment3. KNIME TotalSalesin 10 stores5. KNIME Sales in time for 10 Stores4. R:Promo and Salesper StoreTypeweather csv'sfolder- must be executedbefore loopstart loop-must be executedbefore loopimport i'thcsv fileend loopROSSMANN append weatherget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathrm tempvariablesget StateCodestore_statestate to statecodeget Stateto traintrainROSSMANN append weather.csvget Stateto trainNode 333train_added_weather.csv___L12_meanSales, Cust,SalesPerCustvar:Promo2Openvar:StateHoliday2Openvar:SchoolHoliday2OpenSelect 4 stores:9, 27, 34, 262sort asc by dateNode 344min tempmax tempmean tempsort asc by datefrom x(t) to: x(t), x(t-1), x(t-2), ..., x(t-lag)from x(t) to: x(t), x(t-1), x(t-2), ..., x(t-lag)append allstores toone dfrm first 12M& last 2Mrandom train test split at 80%One last month as holdoutstore id as stringvar:CompetitionTime(in months)to integer()var:PromotionTime(in weeks)replace inString columnthis is just neededfor segmentationROSSMANN segmentation sales data.csvon Store (ID)2.data auditmodel(doesnt takechar vars as input)model Color: AssortmentMean(Sales)in clusterCluster count(apparently sometimesit doesnt evenremember node set up)Sales toCustcolor by clustermore vars availableto see at oncesales to cust per clusterKNIMEk-Means need missing values handledbackup if performance issuesweather data audit5 last months as holdoutNode 434 CSV Reader CSV Reader String to Date&Time Extract Date&TimeFields Math Formula Statistics Flow variables(Parameters) Math Formula Rule-basedRow Filter Rule-basedRow Filter H2O Local Context Table to H2O Linear Regression overlagged (10) values String Manipulation Number To String(deprecated) String Manipulation Math Formula GroupBy String To Number(deprecated) Rule Engine String Manipulation String to Date&Time InteractiveHistogram (local) GroupBy Bar Chart Color Manager Histogram Sales, Cust, Salesper Cust - Charts Store Type and PromotionInfluence Charts List Files Table Row ToVariable Loop Start File Reader Loop End CSV Writer String Manipulation String Replacer String Manipulation String Manipulation String Manipulation String Manipulation Column Filter Joiner File Reader File Reader Joiner CSV Reader CSV Reader Joiner Row Filter CSV Writer Moving Average Math Formula Math Formula Math Formula Rule-basedRow Filter Sorter Row Filter Row Filter Row Filter Row Filter Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Sorter Lag Column Lag Column Sorter Sorter Sorter Concatenate Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Concatenate Concatenate Rule-basedRow Splitter H20 Gradient BoostingMachine Train - Test H20 Gradient BoostingMachine Holdout Number To String(deprecated) Math Formula String To Number(deprecated) Math Formula Rule Engine Write Table: Avg Sales,Cust, SalesPerCust CSV Reader Joiner Statistics k-Means H2O k-Means(deprecated) H2O to Table Color Manager InteractiveHistogram (local) InteractiveHistogram (local) Scatter Plot(local) Scatter Matrix(local) Missing Value Sales Chart - One Store(from Flow Variable) Timer Info CSV Writer Statistics H20 Gradient BoostingMachine Holdout Metanode Data Preparation:Aggregate daily Rossmann sales data into monthly data Data Preparation -Automation of appending data from disparate filesandJoining weather data with sales data Data Preparation:Calculating new variables based on existing column in sales data Data Preparation:Filtering data and selecting 4 store for example analysis Data Preparation:Calculating new variables based on lagged values of weather (+1,+2) and sales and customer data (-12) Data Preparation:Appending together spllited store data and filtering unused in modellingrows Modelling metanodes 2015 prediction Segmentation:Basic segmentation of stores according to their features and aggregated transactional data 1. train1,017,209 rows9 colsstorechange date from default string storage get Year, Month etc from datevar:SalesPerCustomerdata auditcreate flow variablesvar:Cost (as 0.8*Sales+ random() )Select 10 storesexclude Open = 0and Sales = 0CompetitionSince ... DateCompet and PromoPromotionSince ... Datevar:Promo2SinceMonth(formula proxy)AggregateStores to Monthchange StateHolto integerreplace inString columnStateHolidaycreate Date (Monthly)change date from default string storage colored (use Vis Settings for adjustment)Storesand AssortmentColor: Assortment3. KNIME TotalSalesin 10 stores5. KNIME Sales in time for 10 Stores4. R:Promo and Salesper StoreTypeweather csv'sfolder- must be executedbefore loopstart loop-must be executedbefore loopimport i'thcsv fileend loopROSSMANN append weatherget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathget StateNamefrom pathrm tempvariablesget StateCodestore_statestate to statecodeget Stateto traintrainROSSMANN append weather.csvget Stateto trainNode 333train_added_weather.csv___L12_meanSales, Cust,SalesPerCustvar:Promo2Openvar:StateHoliday2Openvar:SchoolHoliday2OpenSelect 4 stores:9, 27, 34, 262sort asc by dateNode 344min tempmax tempmean tempsort asc by datefrom x(t) to: x(t), x(t-1), x(t-2), ..., x(t-lag)from x(t) to: x(t), x(t-1), x(t-2), ..., x(t-lag)append allstores toone dfrm first 12M& last 2Mrandom train test split at 80%One last month as holdoutstore id as stringvar:CompetitionTime(in months)to integer()var:PromotionTime(in weeks)replace inString columnthis is just neededfor segmentationROSSMANN segmentation sales data.csvon Store (ID)2.data auditmodel(doesnt takechar vars as input)model Color: AssortmentMean(Sales)in clusterCluster count(apparently sometimesit doesnt evenremember node set up)Sales toCustcolor by clustermore vars availableto see at oncesales to cust per clusterKNIMEk-Means need missing values handledbackup if performance issuesweather data audit5 last months as holdoutNode 434 CSV Reader CSV Reader String to Date&Time Extract Date&TimeFields Math Formula Statistics Flow variables(Parameters) Math Formula Rule-basedRow Filter Rule-basedRow Filter H2O Local Context Table to H2O Linear Regression overlagged (10) values String Manipulation Number To String(deprecated) String Manipulation Math Formula GroupBy String To Number(deprecated) Rule Engine String Manipulation String to Date&Time InteractiveHistogram (local) GroupBy Bar Chart Color Manager Histogram Sales, Cust, Salesper Cust - Charts Store Type and PromotionInfluence Charts List Files Table Row ToVariable Loop Start File Reader Loop End CSV Writer String Manipulation String Replacer String Manipulation String Manipulation String Manipulation String Manipulation Column Filter Joiner File Reader File Reader Joiner CSV Reader CSV Reader Joiner Row Filter CSV Writer Moving Average Math Formula Math Formula Math Formula Rule-basedRow Filter Sorter Row Filter Row Filter Row Filter Row Filter Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Sorter Lag Column Lag Column Sorter Sorter Sorter Concatenate Lag Column Lag Column Lag Column Lag Column Lag Column Lag Column Concatenate Concatenate Rule-basedRow Splitter H20 Gradient BoostingMachine Train - Test H20 Gradient BoostingMachine Holdout Number To String(deprecated) Math Formula String To Number(deprecated) Math Formula Rule Engine Write Table: Avg Sales,Cust, SalesPerCust CSV Reader Joiner Statistics k-Means H2O k-Means(deprecated) H2O to Table Color Manager InteractiveHistogram (local) InteractiveHistogram (local) Scatter Plot(local) Scatter Matrix(local) Missing Value Sales Chart - One Store(from Flow Variable) Timer Info CSV Writer Statistics H20 Gradient BoostingMachine Holdout Metanode

Nodes

Extensions

Links