Figure 8 - The ADM Main Menu
One of the main tools in Nyamaluma's research wing, used for both research and presenting monitoring feedback to ADMADE communities, is the information and GIS system. Developed in the early days of the project, Nyamaluma's database has enabled thousands of dataforms to be analyzed, and hundreds of tabular and graphical summaries to be generated for community use and applied research. Nyamaluma pioneered the use of GIS software for community-based conservation, and digitized dozens of Survey Department maps for the production of flipchart-sized summary maps of monitoring data. The maps and summaries produced by Nyamaluma have proven to be an effective means of conveying monitoring feedback to community members and catalyzing dialogue around key resource issues. Many community-initiated land use resolutions, that would be considered innovative by any standard, originated at meetings where a group of people gathered around a map showing the locations of hunting revenue, human settlements, habitat, disturbances, animal movements, etc.
As pioneering as Nyamaluma's information system was in the early days, it was constrained by the software and hardware of the early '90s, and its performance was severely limited in a number of regards. By the time this project commenced in 1998, Nyamaluma has acquired newer and more powerful hardware and software. Hence it was a timely opportunity to upgrade the database and take advantage of the newer technology.
CONSTRAINTS
Difficult to operate
The old system used a combination of Lotus 123, dBase IV for DOS, and ArcView GIS. Most of the tabular data was entered through Lotus, a process which was semi-automated with macros, and summaries were produced using Lotus and dBase. Maps and graphical summaries were designed manually in ArcView, using imported summaries that were created in Lotus and dBase. Due to the number of software packages involved, and the difficulty of doing much of the work manually, only a very small number of staff could operate the system, and only the technical advisor could make any real significant changes to the design or structure of the data. Operating the database was also time intensive, as only a small part of the data processing was automated.
Restricted to single-area, single-year summaries
In the original system, the files for each Unit were saved in separate directories, and additional sub-directories created for each year. For the tabular data, each year was saved on a separate worksheet and each dataset was saved in a separate workbook. The spatial data was also divided into separate coverages for each Unit. While this file structure was useful in keeping data organized, it made conducting analyses across years or across multiple GMAs tedious almost to the point of almost being impossible. While this was not a major constraint in providing data summaries for individual units, it made seeing the 'big picture' rather difficult.
Little error checking
Due to the limitations of spreadsheets, error checking depended heavily on the skills and experience of the data entry clerk. Problems such as inconsistent units of measurement, inconsistent spelling of names, incorrect dates, and occasional outliers limited the reliability of certain types of analyses. For example, it was difficult to get an accurate count of all staff who worked in a Unit over time because some names appeared twice under different spellings, while others could have just been copied over from the previous year without further inspection, etc.
Unwieldy file system
Maintaining the files in the old information system was an administrator's nightmare. The hundreds of directories, sub-directories, and files made copying or backing up the database challenging. A more severe constraint that may have impacted data quality was the use of multiple computers for entering data, making it incumbent upon the technicians to keep track of which files on which computers are the most up to date. Updating files on other machines carried a risk that more recent data could be overwritten with older data.
Difficult to expand
Adding new datasets into the information system was challenging, because new files had to be integrated into both the directory system, data structure, and multiple software formats. Creating new summaries of data was equally difficult, because summaries had to be either done manually or by programming new Lotus macros. This constraint on the system's expansion and flexibility not only affected staff time, but also the number of datasets that could be entered. Very important datasets, including all field patrol observations, observations from safari hunts, were not entered into the old information system at all because of the limitations of the software. Other datasets, such as poacher case records, and crop damage, were entered but not integrated with other data, so could not analyzed against other variables.
Map production was manual
As mentioned previously, one of the most important outputs of Nyamaluma's information system were the flipchart-sized maps of monitoring data that were returned to communities. Creating these maps in ArcView, although flexible, involves a complicated series of steps that requires a significant amount of training and many hours of staff time.
Data never left Nyamaluma
Because it was difficult to extract data except in the limited preset formats, and the clunky file and software system made it impossible to share data electronically, Nyamaluma's research unit struggled to meet the information needs of its many external stakeholders in Zambia and abroad. Among the donors and wildlife sector, Nyamaluma developed a reputation of being miserly with data, failing to share results with even its closest institutional partners.
OBJECTIVES FOR A NEW INFORMATION SYSTEM
Integrate datasets
Many of the constraints of the old information system stemmed from the disjointed data and file structure. Hence a key objective of the new system was to integrate the major datasets under one relational structure. In other words, all datasets for all years and all Units should be combined together. This would enable producing summaries and analyses based on data from different years and/or GMAs.
Make user-friendly
It was desired that any new database system should be a lot more user-friendly. This would allow a greater number of Nyamaluma staff to input data and generate outputs, and would minimize the amount of disruption in the program resulting from transitions in technical staff. A user-friendly interface would also decrease the likelihood that certain types of errors would occur, and allow the highly capable technical staff to spend less time on repetitive tasks and focus more on the analytical side of data management.
Improved error checking
Enforcing data integrity and developing built-in error checks were other desired features for the new database system. The strategy of storing the names of Units, species, scouts, etc. with identification numbers instead of text strings is one example of a strategy that can reduce potential data errors. Other desired error checking features included automatic checking for numbers which should fall within a certain range (e.g., trophy sizes, dates, etc.), and ensuring that records are not entered more than once.
Automate standard outputs & analyses
To save staff time and improve reliability, there was a need to automate many of the standard outputs of the database, including tabular summaries, maps, and charts. Automating the standard analyses also improves the consistency of the outputs produced. One of the weaknesses of the old system was that the manually created maps and charts often used different color schemes, column headings, layout design, etc. Although such variations might be insignificant to people who are well trained, they can disorient rural people who may be less visually literate.
Provide multi-user features
More and more of the staff at Nyamaluma are using computers in their work, including some extension teams who even carry laptops with them on field visits. One of the hopes for this new database was that monitoring data could be entered in the field, and summaries provided to communities immediately, eliminating the long feedback delay caused by transport to and from Nyamaluma. Hence the new database had to provide multi-user features such as the ability to synchronize datasets across non-networked computers. Another one of the troubles with the old system was that there were normally multiple copies of the same files on different machines, and it was left to the computer staff to remember which was most recent.
Facilitate future expansion
An information system which supports a program as dynamic as ADMADE needs to be able to grow with the times. New datasets, new summaries, new layers of spatial data, new maps, and new users, are all examples of likely changes the database will need to accommodate over its life span.
Improve documentation
Documentation was an important component for many of the desired features in the new database, including user-friendliness, multi-user features, and expansion. Nyamaluma's first information system was fairly well documented in a technical manual, although many of the 'tricks' were only acquired through experience, and expansion of the system was not a topic in the manual. Documentation for the new database required not just the standard printed materials, but also context sensitive online help. Furthermore, since the new database was going to have multiple architects and an order of magnitude greater number of ready-made summaries and outputs, there was a need to integrate documentation for the individual outputs. In other words, every tabular summary, map, and graph needs a mechanism to allow the user to find out what the summary is trying to capture, how it is calculated, and who created it.
Capture qualitative data
Information systems are most adept at capturing and analyzing quantitative data, however in programs as complex as CBNRM projects much of the most interesting data and almost all of the interpretation can only be expressed qualitatively (i.e., through text or pictures). In addition to the substantial use of dataforms, which are designed to capture quantitative observations, we should not forget that ADMADE's research unit has amassed a significant quantity of qualitative data, much of it written down in the form of field reports, correspondence, trip plans, workshop proceedings, and land use resolutions. This qualitative data is critical to interpreting the quantitative results of monitoring data, and also needs to be available to the user through a common interface.
Make accessible to wider audience
Although it will be some time before rural communities will have the capacity to use computers, there are a host of other potential users for ADMADE's monitoring database. These are described more fully in Chapter 2, but some of the main ones include senior officers in ZWA, USAID, and ADMADE's institutional partners within Zambia. With a cleaner, more robust information system, ADMADE for the first time could have the technical capacity to share all or some of its raw data or summaries with other institutions, either electronically or in hard copy. There are numerous considerations to regard when sharing raw data, it was desired that the new database should at least remove some of the technical hurdles that had been hampering ADMADE's ability to disseminate monitoring results.
The new database, called the ADMADE Data Manager (ADM), is now the working information system at Nyamaluma and is also being used at the ADMADE coordinating office in Chilanga and USAID in Lusaka. ADM uses a combination of built-in Access features and customized enhancements, explained below.
BUILT-IN ACCESS FEATURES
Relational data structure
Like most modern databases, Access supports relational data structures, which simply means that related data is divided into different tables. For example there is a table for staff, a table for species, a table for Units, a table for field patrols, etc. The information in all of these tables are linked together with ID numbers, which computers can process much faster than text. Using a relational data structure saves a significant amount of disk space and improves performance when querying or summarizing data.
Enforced data integrity
Access is well equipped to ensure that data saved in related tables does not violate referential integrity rules, and that all required fields are filled in. When data integrity is enforced, it becomes impossible to add the same record twice (in most cases), and impossible to enter incomplete data. For example, you couldn't enter a new field patrol record unless there was a valid value for the date it took place and Unit where it originated from. This feature, along with the relational data structure, eliminates many of the potential errors that can be caused by inconsistent spellings, partial records, etc.
Replication
Replication is another built-in feature of Access that allows multiple copies of the same database to communicate with each other and synchronize the data. Replication is a real lifesaver in a facility like Nyamaluma, where two or three copies of the database are needed just for data entry workstations, and others may be needed for performing analyses and generating outputs. With replication, it is relatively simple to make certain all copies of the database are using the most up-to-date data, and each copy has the latest preset summaries and outputs.
CUSTOM DESIGNED FEATURES
User-friendly menu system. A database with as many different types of datasets and summaries as ADMADE requires a menu system to navigate among the many different choices. ADM features a standard three-tiered point-and-click Main Menu, and a simple single document interface, which means you only see one window at a time. Choices on the Main Menu can be easily expanded or modified using the Menu Manager. The menu system also features integrated object filtering and documentation, which are described below. See Appendix IV for a listing of the choices on the ADM Main Menu (as of May1999).
Figure 8 - The ADM Main Menu
Figure 9 - The ADM Filter Manager
Figure 10 - The ADM About window - part of the integrated object documentation
Integrated mapping capability
Using MapObjects to serve as the link between the tabular and spatial data, ADM includes user-friendly Interactive Maps for visualizing the spatial element of monitoring data. These maps present spatial summaries of data, such as an annual summary of which grids are used by safari clients, where poaching activity has been observed, which areas generate the most revenue, field patrol effort, etc. Interactive Maps have many of the same toolbar options as ArcView, including the ability to add labels, change display colors or the classification scheme, make additional layers visible, create a legend, pan and zoom, etc. New maps can be easily created and added to the menu system by entering a new map definition, which specifies properties such as the layers which should be added, data the map should be linked to, etc. All interactive maps can be printed, copied to the clipboard, or sent to PowerPoint, and all make use of the standard features of the menu system, including filtering with the Filter Manager and plain-English documentation.
ADM's interactive maps use many of the spatial layers that have been digitized at Nyamaluma, as well as a few others collected from various sources. All layers are national in scope, allowing maps to be produced of multiple GMAs. The following GIS layers are available:
| Hunting blocks* Units* 5km2 grids* Scout camps* GMA roads* GMA rivers* National parks Districts Provinces | Provincial capitals National roads National rivers Railroads Utility lines Airports Villages Wetlands |
*digitized at Nyamaluma
Poster-size layouts. One of the most important outputs of Nyamaluma's database are the large poster-sized layouts of monitoring data, which are used in community meetings and workshops. ADM uses a programming technique called OLE automation to automatically create new layouts using PowerPoint using data from Access. PowerPoint is the Microsoft graphics-presentation application that also comes with Office 97, and contains a number of drawing tools and commands. Layouts that are created by ADM using PowerPoint can contain any combination of maps, charts, text, or summary tables. They can create in 3-4 minutes what previously took a trained technician a couple of hours to produce, a significant time savings when multiplied by 15-20 Units two or three times a year. New layouts can be designed by creating a "Slide Template" definition, which specifies the layout of different elements on the page. Users can use the Filter Manager to select which data should be summarized when creating the layout. Once created in PowerPoint, users can adjust the layout, add other elements such as digital photographs or clipart, and finally print it out on a color printer or plotter.
Figure 11 - ADM automates the creation of poster-sized printouts that combine maps, graphs, and tables
Documentation
A comprehensive Users Guide has been written describing how to use ADM. The Users Guide has sections both for novice users as well as technical staff who need to know how to maintain and expand the system. The Users Guide comes both in printed format as well as a context-sensitive Windows help file (which means that the 'Help' buttons in ADM will take you the correct section of the Users Guide). A complete copy of the ADM Users Guide, as well as a shorter overview guide, can be downloaded from http://nersp.nerdc.ufl.edu/~alyons/zm/adm.html.
The web is an appropriate medium for ADMADE to use for several reasons. Using the web resolves one of the biggest hurdles ADMADE has faced in disseminating information: the difficulty and expense in producing and distributing hard copies. Nyamaluma, where the vast majority of monitoring and publicity materials are produced, is quite remote and only has one reliable photocopy machine. In addition to not wanting to overwork the photocopier, paper and toner are in short supply and expensive to buy locally. Transporting hardcopy documents is expensive and logistically complicated. Zambia is a large country where transport costs are relatively expensive. Creating enough copies of every report, newsletter, educational or publicity material for all potentially interested stakeholders is simply not feasible for this shoe-string program.
The web removes many of the barriers of distributing information. Although some of ADMADE's stakeholders do not have access to the internet, in particular local communities, many of the national and international stakeholders do have web access. Zambia is fortunate to be relatively well 'wired' into the internet, with at least one national ISP and dial-in services in many parts of the country. ADMADE can all but eliminate the costs and technical barriers to distributing information to these stakeholders.
ADMADE is also particularly well placed to share data electronically because most of its program and monitoring materials are already prepared in electronic format. Nyamaluma pioneered the use of database and GIS technology for CBNRM monitoring, and has been using word processing, digital photography, and desktop publishing software for years to prepare education materials, newsletters, and reports. Converting these materials to a format suitable for the web is technically trivial with commonly available software tools. Even Nyamaluma's recently upgraded monitoring database could be put online, either statically or interactively.
To develop the web site, a list of the potential audiences was first developed. These included members of the wildlife sector in Zambia, international conservation organizations, tourists, safari hunters, academics, donors, and the general public. Secondly, a list of thematic menu sections was developed. These included introductory material about ADMADE, wildlife conservation accomplishments, community development, safari hunting, Nyamaluma, monitoring and GIS, publications, bibliography, and related links. Finally pages were designed for each of the sections, largely using text and images from existing documents.
ADMADE's web page made its debut in January 1999. Although it still has a long way to go, the general public is now able to get first hand information about ADMADE, its mission, strategies, and results.
In order to strengthen monitoring capacity at the Units, Nyamaluma scheduled an advanced course for village scouts and deputy unit leaders, about ¾ of which focused on topics related to monitoring. This one-week workshop was held in May 1999, and was attended by 44 scouts from all ADMADE areas. The workshop objectives related to monitoring included:
Monitoring will be covered in other courses scheduled for 1999. A workshop for Unit leaders held in June addressed monitoring issues, and various workshops for the newly elected Community Resource Boards also touched on monitoring. ADMADE will need to continue to provide training, both at Nyamaluma and in the field, in order to enable Unit staff and communities to analyze their own monitoring data for exercises such as quota setting, setting work targets, and resolving land use conflicts.