Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations WHITE
Tabe of Contents Executive Summary... 3 Chapter 1 - Datawatch Visuaizations... 4 Chapter 2 - Snapshot Visuaizations... 5 Bar Graph... 5 Buet Graph... 6 Cross Tab Pivot Tabe... 6 Dot Pot... 6 Heat Map...7 Heat Matrix... 8 Map Pot... 8 Numeric Line Graph... 9 Numeric Neede Graph... 9 Numeric Stacked Neede Graph... 10 Pie Chart... 10 Scatter Pot... 11 Geographic Scatter Pot... 11 Shapes/Choropeth... 12 Surface Pot... 12 3D Surface Pot... 13 Treemap... 13 Chapter 3 - Time Series Visuaizations... 14 Candestick Graph... 14 Horizon Graph... 14 Line Graph... 15 Neede Graph... 15 OHLC Graph... 16 Percentage Area Graph... 16 Spread Graph... 16 Stack Graph...17 Stacked and Grouped Neede Graph...17 Time Series Scatter Pot... 18 Time Series Combination Graph... 18 Chapter 4 - Mixed Mode Visuaizations... 19 Tabe... 19 2
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Executive Summary Visuay Design, Discover and Expore New Insights from Any Data in Rea Time Visua Data Discovery In today s market, businesses must everage every shred of data to stay competitive. And with the amount of data continuay growing, it s critica for organizations to anayze, understand, and interact with this data regardess of its type (variety), its size (voume) or the speed in which it is deivered (veocity). Visua Data Discovery is critica if you are truy ooking to get more from your data and not just graph something where you undoubtedy aready know the answer. The Datawatch Desktop soution ets you quicky start asking questions to see hidden patterns, spot probems and identify missed opportunities without programming or scripting. Anayze Data in Motion and at Rest The idea of rea time can be very confusing, since virtuay a software companies that do visuaizations say they can hande rea-time requirements. However, what most of them mean is that their software goes out and requests an update from an externa data source every time fresh data is needed. The data is accurate as of that exact moment, but it becomes out of date immediatey since no further updates are avaiabe unti a compete refresh is done. The information that is being queried comes to rest in a rea-time data warehouse or database. Datawatch is unique in its abiity to visuaize data in motion by consuming data streams from sources ike CEP engines and message brokers which constanty push information into the system instanty, as it happens, on a tick-by-tick basis. So you can see exacty what is happening, as it happens and know precisey how your business is performing. Extract and Transform Data from Existing Reports Data is diverse and rarey presents itsef in a form perfecty structured and ready for anaysis. Some of the most vauabe information inside your organization is ocked in static operationa reports that provide a necessary and trusted set of information but is infexibe. You aso have critica data that come from outside the four was of your company (ike invoices, statements, forms, market data) where you don t have access to the underying systems. Datawatch aows users to access, extract and transform any static data into ive data for visuaization, anaysis and sharing with other users and systems. Without programming, a business user opens the report or fie in Datawatch and can point and cick on the data to be extracted. Now, et s get started. Datawatch supports a wide range of visuaizations to give you an easy and fast understanding of a of your data, however, there is no one visuaization that is idea for every purpose. This paper outines and provides guidance for choosing the appropriate visuaization for the right anaytica task at hand. 3
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations CHAPTER 1 Datawatch Visuaizations Datawatch Desktop software supports a wide range of information visuaizations, incuding our we-known Treemaps, Heat Maps, Scatter Pots, Horizon Graphs, and a wide range of other great visuaizations designed for fast comprehension and easy interpretation of static, time series, rea-time streaming, and historic data sets. As no one visuaization is idea for every purpose, the appropriate visuaization for the anaytica task at hand must be used. Here are some genera recommendations: Anaytica Task Read numeric vaues quicky Performance against a KPI Performance across a singe variabe for a sma number of data eements, with different magnitudes Performance across a singe variabe for a sma number of data eements, each with simiar magnitudes Performance across a singe variabe for a arge number of data items Performance across a singe variabe for a arge number of data items, which have different importance vaues Performance across a hierarchica or grouped dataset Correation between two categories of data Correation between two or more numeric data coumns Geographic correations of data Correation over both a singe numeric data coumn and various categories of data Trending performance across ordered categories Trending performance between two numeric variabes Recommended Visuaization Tabe / Pivot Tabe Buet Graph Bar Graph Dot Pot Heat Map Treemap Treemap Heat Matrix Scatter Pot Map Pot Geographic Scatter Pot Dot Pot Dot Pot Numeric Line Graph Trending performance between three numeric variabes Surface Pot (and 3D) Trending performance across time Time based Ranking Time Based Contributions Time Based Correations between time series Time Based Transactions Financia Time Series Distributions Auction Price & Interest/Voume Distribution Geospatia Area Densities Spread between two time series Line Graph Line Graph with Ranking Axis Stack Graph Horizon Graph Neede Graph Cande Stick or OHLC Graph Numeric Neede Graph Shapes Spread Graph 4
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations CHAPTER 2 Snapshot Visuaizations Some of the most common use cases for data visuaization software require the system to dispay information about a data set as it exists at a particuar point in time. These snapshot visuaizations are extremey usefu for understanding reative quantitative and quaitative measures and enabe users to gain a comprehensive understanding of very compex data sets very quicky. Figure 3-2. A standard bar graph. Bar Graph Bar Graphs are probaby the best known visuaization for quantitative data. You can dispay Desktop Designer Bar Graphs either horizontay or verticay. These graphs are avaiabe in three variants: Figure 3-3. A grouped bar graph. Standard Grouped Stacked In each case, you can sort the ayout of the bar graph according to your requirements, and, with hierarchica data, the graph represents the netted position at each aggregated depth eve. You can aso use the Bar Graph visuaization to dispay demographic data in so-caed Tornado Charts or Popuation Pyramids. Figure 3-4. A stacked bar graph. Figure 3-1. A horizonta bar graph. Figure 3-5. A stacked bar graph showing a tornado chart ayout. 5
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Buet Graph Buet Graphs were designed by Stephen Few to remove unnecessary cutter and instead focus on visuaizing metrics ike Key Performance Indicators (KPI). Research has shown that Buet Graphs are easier to interpret in ess time than the radia gauges or speedometers often seen in BI dashboards. Figure 3-6. A horizonta buet graph. Figure 3-9. A pivot tabe with intense coors. Dot Pot Dot Pots have two primary use cases: A more effective aternative to a Bar Graph Figure 3-7. A vertica buet graph. Cross Tab Pivot Tabe Athough seected visuaizations can be cross*tabbed into sma mutipes, each showing subsets of the origina data set, the Cross Tab can itsef be used to dispay a Pivot tabe. A distribution dispay simiar to a Scatter Pot Dot Pots are an effective aternative to Bar Graphs, particuary in cases where the data being anayzed contains many simiar numeric vaues. In comparison with the Bar Graph, Dot Pots do not use a zero baseine and are ess cuttered. This makes it easier to add additiona data variabes to the visuaization. Pivot Tabes support a singe numeric vaue being represented at the cross point of hierarchica rows and coumns. Each intersection ce can dispay the aggregated numeric vaue and its associated coor range, which can be subdued or intense. Numeric abes can aso be removed to produce a Heat Matrix. Figure 3-10. A sampe horizonta bar graph showing revenues versus forecasts. Dot Pots can aso be used to represent data distributions in which one axis is numeric whie the other axis is categorica. Figure 3-8. A pivot tabe with subdued coors. 6
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Scatter Pots using such data sets can be misinterpreted; Dot Pots of the same data are unambiguous and easy to understand. Heat Map A Heat Map is a specia type of coor-based data visuaization that is we suited for anayzing arge fat data voumes using an intuitive graphica dispay. Heat maps are good at representing arge numbers of data points in ways that woud be unwiedy and hard to interpret using traditiona tabes or charts. A Heat Map represents each item in the data set as an equay-sized ce, unike a Treemap that uses the size of the box to represent a quaitative vaue and ocation to represent hierarchica reationships. In a Heat map, the coor of the square represents a quantitative vaue reative to the other boxes in the Heat map, whie the ocation can represent the sorting of another quantitative or categorica vaue. This aows the anayst to see a of the data items simutaneousy. The user can aso hover over any item to bring up more detaied information on demand. Figure 3-11. A dot pot of revenues versus forecasts Figure 3-14. A heat map without sorting. Figure 3-12. A distribution dot pot. Figure 3-13. A business period categorica ine graph Figure 3-15. A heat map with sorting. 7
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Heat Matrix A Heat Matrix is simiar to both the Heat Map and Treemap in that it dispays many different data items and represents the vaue for each item using coors. However, unike its cousins, the Heat Matrix has a defined structure where two data attributes define each axis, thus producing a correation matrix. Within the Heat Matrix, each coumn and row represents a unique attribute, and the point where two items intersect represents a unique combination of the two attributes. The matrix can dispay abes within each intersecting tie or simpy dispay coor. Our Heat Matrix data visuaization heps our cients identify correations within their data sets using an intuitive graphica dispay. Map Pot Use Map Pots to dispay geographic data, where you have ongitudes and atitudes associated with individua data points. These pots ceary show data correations and custering that is geographic in nature. In a Map Pot, the visuaization expects Latitude and Longitude measures to be associated. It wi then retrieve from the seected map tie provider the appropriate background map to dispay under the data points. This background map is constructed by retrieving individua map ties at set zoom eves. As the background map is provided automaticay, it reies on: a) A range of suppied ongitudes & atitudes to provide a bounding area. b) An active Internet connection to retrieve the map tie images. Datawatch ships with a number of cross reference datasets to determine the appropriate atitude/ongitude for datasets. These have been provided through subsets of the data avaiabe at GeoNames.org. ( http://www.geonames.org ) More detaied geo-coding data is avaiabe from this website, and many others. Figure 3-16. A heat matrix. Zooming into the map wi cause, new map ties to be retrieved, and a new background map image behind the data points to be dispayed. For exampe increasingy zooming into Northern Europe woud produce: Figure 3-17. An FX cross rates heat matrix. Figure 3-18. Map Pot of each country gobay. 8
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Numeric Line Graph Numeric Line Graphs differ from the standard ine graph in that they have a numeric X axis, rather than one based upon time. They are commony used in both scientific and financia scenarios to show trends in functions that are based on two numeric inputs (X and Y). Figure 3-19. Map Pot of the northern hemisphere. Common uses incude the dispay of Yied Curves. Numeric Line Graphs can aso be used to dispay seected cuts through a Surface Pot. Figure 3-20. Map Pot of Europe. Figure 3-23. A numeric ine graph. Numeric Neede Graph Numeric Neede Graphs dispay price distributions. Figure 3-21. Map Pot of the Netherands, Begium and surrounding countries. Unike a traditiona Bar Graph, the X Axis is numeric rather than categorica. Bars are positioned aong the X axis according to their X vaue, and their height is determined by their Y vaues. This aows gaps, and custering in price to be more accuratey identified. Figure 3-22. Map Pot of Antwerp and the surrounding area Figure 3-24. A numeric neede graph. 9
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Numeric Stacked Neede Graph Numeric Stacked Needes again dispay price distributions. Unike the standard Numeric Neede Graph, mutipe items can be identified at a singe price. A common usage is dispaying cient activity within an order book. Figure 3-27. A typica pie chart. Figure 3-25. A separated numeric stacked neede graph. Figure 3-28. A mutieve pie chart (Sun burst). Figure 3-26. A numeric stacked neede graph. Pie Chart Pie Charts are one of the odest and best known visuaizations for dispaying contributions to a tota. Desktop Designer can produce standard Pie Charts in which the pie sice represents a numeric variabe that is proportiona to the tota size of the pie. The coor variabe can represent either a category or another numeric variabe. Figure 3-29. A mutieve pie chart with deeper hierarchy. Pie Charts can be fat, showing a singe set of sices. They can aso be hierarchica and dispay mutipe eves of data in a variant caed a Mutieve Pie Chart. This is aso known as a Sun Burst or a Radia Treemap. The user can modify the visibe depth eve and dri into particuar sices to investigate further detai. The center of a mutieve Pie Chart can be cut to form a Donut Chart. However, rather than simpy eaving the area bank, Datawatch Designer Pie Charts show the aggregate coor for the compete data set Figure 3-30. A mutieve pie chart with the center showing aggregate coors. 10
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Scatter Pot Scatter Pots are used to identify trends, custering and outiers across a number of numeric variabes, especiay when investigating arge data voumes. Each scatter point is represented by: x Position Y Position Size Coor (numeric or categorica) A ine of best fit can aso be added to highight outiers. Desktop Designer s Scatter Pot data visuaizations are easy to set up and highy customizabe. You can configure your dispay in ways that wi make the most sense to you and your users, and users have a the toos they need to fiter and manipuate the Scatter Pot to concentrate on the most reevant subsets in the data. Geographic Scatter Pot Use Geographic Scatter Pots to dispay data where physica ocation is important, and the background map image can be manuay provided. These pots ceary show data correations and custering that is geographic in nature, and typicay used for non-standard mapping. If a standard map is required then it is ikey that the Map Pot shoud be used instead. In Geographic Scatter Pots, the X and Y coordinates can correspond to ongitude and atitude. The coor and size of each scatter point represent other data variabes. As with standard statistica Scatter Pots, you can zoom and pan within the visuaization to focus on specific areas of interest, but the underying map image wi not change. As the background map image is manuay provided, the visuaization can be used for non-traditiona maps, such as interna foor pans. Figure 3-33. A geographic scatter pot. Figure 3-31. A scatter pot with ine of best fit. Figure 3-32. A scatter pot with square scatter points. 11
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Shapes/Choropeth The Shapes visuaization aows the dispay of Choropeth Graphs and other dispays buit from SVG Paths. The Shapes visuaization can be used to dispay data where both physica ocation and size are important. They ceary show data correations and custering that is geospatia in nature. Unike the Geographic Scatter Pot, the size of each shape is fixed, imparting the importance of the item. As a consequence, data shoud be reative to each shape size, such as area densities. Figure 3-35. A surface pot with stepped coors. Figure 3-34. A shape visuaization. Figure 3-36. A surface pot with continuous coors. Surface Pot Surface Pots are used to identify trends and outiers within numeric surfaces. The Surface is made up of a series of points where each point has: x Position Y Position Coor (which represents the Z axis). The Surface Pot can support data sets where the X and Y positions can both be reguar and irreguar in their distribution. Additionay, the coor scae can be continuous or stepped to show a surface gradient. 3D Surface Pot 3D Surface Pots are a 3D perspective version of the 2D Surface Pot. They provide a cearer understanding of the overa shape of the surface but they aso introduce occusion probems; not a data points can be seen due to the dispay perspective. The Surface Pot 3D is made up of a series of points where each point has: x Position Y Position z Position (encoded by coor) The Surface Pot 3D can support data sets where the X and Y positions can both be reguar and irreguar in their distribution. The coor scae can be continuous or stepped to show a surface gradient. 12
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Grid ines, a ground pane, and markers for data points can be shown if required. Our Treemaps are not static pictures. The rea vaue of the visuaization is quicky apparent when you interact with the data. Users can zoom, fiter, and view detais on demand, as we as ink to and highight other sources of information. For exampe, fund managers can ink to a trading system directy from within the Treemap. EX supports three different styes of Treemaps: Cassic Treemaps Windows Treemaps Custer Treemaps Figure 3-37. A 3D surface pot with stepped coors. Figure 3-39. Cassic Stye Treemap Figure 3-38. A 3D surface pot with continuous coors. Treemap Treemaps represent hierarchica data sets, showing both each eve in the hierarchy and how they interact with each other. They are represented by a coorfu mosaic of rectanguar ces based on your data. The size of a ce refects its importance. The coor conveys urgency or variance: Figure 3-40. Windows Stye Treemap White Target/Benchmark Performance Red Under Performance Bue Over Performance The intensity of the red or bue shades indicates the eve of under- or over-performance. Most peope can earn to understand the information presented in a Treemap in under a minute even if that Treemap is showing data representing an underying data set of thousands of records. Figure 3-41. Custer Stye Treemap 13
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations CHAPTER 3 Time Series Visuaizations The abiity to hande very arge quantities of mutivariate time series data is an essentia eement in a compete visua anaysis system. Desktop Designer offers a range of speciaized data visuaizations, incuding Horizon Graphs, Stack Graphs, and Line Graphs, designed specificay to make anayzing historica data easier and more efficient. The software s abiity to connect to traditiona row-oriented reationa databases or coumnoriented databases is key to supporting fast, responsive muti-dimensiona anaysis of arge data sets. Our time series capabiities are especiay important for users in goba investment banks, hedge funds, proprietary trading firms, and exchanges. Horizon Graph Horizon Graphs are a fantastic way to overview a arge number of time series in a imited rectanguar space. Since this visuaization packs the information in a ine graph in 1/6th the space through smart pre-attentive coor encoding, it aows for an overview of a arge number of time series. Users can scan huge amounts of data points across a reevant time series and immediatey identify areas of concern that require coser scrutiny. Our Horizon Graph visuaization is particuary usefu when you need to see a arge number of time series on a singe screen. This makes it easy to compare trends and spot patterns that woud be very difficut or impossibe to see in a standard report. Candestick Graph Candestick graphs are a traditiona financia visuaization for dispay of time-based price distributions. Specificay, for each time sice, they dispay: Opening Price Highest Price Lowest Price Figure 2-3. A horizon graph. Cosing Price The Cande is fied if the cosing price is ower than the open and empty if the cosing price is higher than the open. The vertica ine (or cande wick) dispays the range of traded prices across the period. Figure 2-4. Starting with a ine graph. Figure 2-1. A cande stick graph. Figure 2-5. Creating coor performance bands. Figure 2-2. Cose up of data points in a cande stick graph. 14
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Neede Graph Neede Graphs dispay time-based transactions or occurrence frequencies, rather than time-based trends. They are simpy time-based Bar Graphs where each bar is ocated at a particuar time point on the axis. They work especiay we when combined with a Line Graph. Figure 2-6. Inverting negative regions. The most common use of a Neede Graph is when showing the trading voume for a stock, typicay underneath the price performance. Figure 2-7. Coapsing the performance bands. Figure 2-10. A neede graph. Figure 2-8. A horizon graph is finay created. Line Graph Line Graphs are easy to understand and are a great way to communicate important time-based trends, custering, and outiers. They work especiay we when comparing ten or fewer data sets (our Horizon Graph is a good soution for dispaying time series data with ten or more data sets). Figure 2-11. A neede graph in combination with a ine graph. Figure 2-9. A ine graph. 15
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations OHLC Graph OHLC Graphs aso dispay time based distributions of price data. For each time sice, they dispay: Opening Price Highest Price Lowest Price Percentage Area Graph A Percentage Area Graph is ike a Treemap spread out over time; you can see how each constituent part contributes to the tota at any point in the time series. It is an exceent choice for visuaizing time series data when you are interested in seeing the reative contributions for each data set in the series, regardess of the absoute tota. Cosing Price Simiar to the Candestick Graph, a vertica ine defines the range of traded prices across the period. However, in this case, the eft notch determines the opening price and the right notch determines the cosing price. Figure 2-12. An OHLC graph. Figure 2-14. A percentage area graph. Spread Graph The Spread Graph dispays the variance or spread between two time-based data series. Typica use cases incude comparing a stock s price performance to an Index or a bond s yied to a benchmark rate. Figure 2-13. Cose up of data points in an OHLC graph. Figure 2-15. A spread graph with inear interpoation. Figure 2-16. A spread graph with stepped interpoation. 16
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Stack Graph Stack Graphs et you visuaize quantitative changes to severa data sets over time, and you can see how each data point contributes to the tota. As with the Treemap the Height of the stack reates Importance, whie the coor reates Urgency or variance. Stack Graphs are a great way to ook at revenue or gross profit figures over time across severa product ines. Stack Graphs are aso good to use when you have up to ten or eeven time series data sets to ook at, especiay for data sets that have a arge number of data points. Figure 2-17. A portfoio stack graph. Figure 2-19. Sampes of stacked and grouped neede graphs. Figure 2-20. Stacked neede graph (Turnover by exchange). Figure 2-18. An oi production stack graph. Stacked and Grouped Neede Graph Stacked and Grouped Neede Graphs dispay time-based transactions or occurrence frequencies, simiar to the standard Neede Graph. Figure 2-21. Stacked neede graph (Buy and se voume). It aows each transaction to be spit into its components, aowing contributions to the tota to be viewed across time. Common uses incude spitting of transaction voumes by venue or by direction (Buy/Se). 17
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations Figure 2-22. Stacked neede graph (Buy and se voume - separated). Figure 2-25. A time series scatter pot (Trade voume with best bid and offer). Time Series Combination Graph Figure 2-23. Stacked neede graph (Net voume through aggregation Buy/se). Time Series Scatter Pot Time Series Scatter Pots dispay time-based transactions, simiar to the Neede graphs. Like the scatter pot, it dispays individua data points (or transactions), with a given numeric Y vaue and a given timestamp X vaue. Common uses incude dispaying transaction voume across time reative to the price at which the voume was executed. Typicay, the graph is combined with ine graphs to show the scatter points reative to defined boundaries. Figure 2-25 shows trade voumes and prices reative to the best bid and offer across time. The Time Series Combination Graph, combines a series of time series visuaizations as individua ayers of the tota dispay. As a consequence more compex time series visuaizations can be buit from the base visuas. Each visua can be assigned to either the eft or right Y axes, aowing mutipe scaes to be represented. For exampe the foowing visuaization incudes: Cande Stick Graph Showing the distribution of prices (OHLC) Line Graphs Showing moving averages of the cosing price Neede Graph Showing traded voume across the period Spread Graph Showing a price band across the period. Each of the visuas has a defined Z order, which in this case paces from back to front: Spread, Neede, Cande Stick, Line Figure 2-24. A time series scatter pot. Figure 2-26. A time series combination graph. 18
Understanding the Vaue of Visua Data Discovery A Guide to Visuaizations CHAPTER 4 Mixed Mode Visuaizations Mixed Mode Visuaizations are capabe of dispaying time series or snapshot data. In some cases, these types of visuaizations can dispay both time series and snapshot data simutaneousy Tabe A tabe can be used to dispay a sma dataset where a the vaues are visibe or the aggregate vaues of a arger data set. Figure 4-1. A simpe tabe. The tabe can be configured to show hierarchies, aowing sub totas and grand totas to be dispayed. Additionay, branches of the hierarchy can be expanded and coapsed. The tabe can be sorted by cicking on a coumn heading, and sorting is appied across the defined hierarchy. Coumns ces can be represented in their vaue form or, aternativey, graphicay as a series of micro-charts incuding: Buet Graph Bar Graph Dot Pot Figure 4-2. A tabe with hierarchy, totas, and microcharts. Line Graph Figure 4-3. A tabe showing Snapshot and time series trends About Datawatch Corporation Datawatch Corporation (NASDAQ-CM: DWCH) provides visua data discovery software that optimizes any data regardess of its variety, voume, or veocity deivering next generation anaytics to revea vauabe insights for improving business. Its unique abiity to integrate structured, unstructured, and semi-structured sources ike reports, PDF fies and EDI streams with rea-time streaming data into visuay rich anaytic appications aows users to dynamicay discover key factors that impact any operationa aspect of their business. This abiity to perform visua discovery against any data sets Datawatch apart in the big data and visuaization markets. Organizations of every size, wordwide use Datawatch products, incuding 99 of the Fortune 100. Datawatch is headquartered in Chemsford, Massachusetts with offices in New York, London, Munich, Stockhom, Singapore, Sydney and Mania, and with partners and customers in more than 100 countries wordwide. See the Whoe Story for yoursef by downoading the free tria at www.datawatch.com/tria. Contact Datawatch For more information about Datawatch, contact us directy, www.datawatch.com, +1 978.275.8222, or saes@datawatch.com. Datawatch Corporation, 271 Mi Road, Quorum Office Park, Chemsford, MA 01824, USA To free US (800) 445.3311 or +1 978.441.2200 2014 Datawatch Corporation. A rights reserved. Datawatch, the Datawatch ogo and other product names, ogos, and tag ines are trademarks of Datawatch Corporation. A other trademarks or registered trademarks are properties of their respective owners. 19