Since its launch in 2009, JUSP has played an increasingly valued role within UK Higher Education
By bringing together journal usage data for more than 70 publishers into a single place, JUSP creates huge economies in terms of staff time and costs across the sector. In this article, JUSP team member Paul Meehan describes a typical day working within the team
It’s 9:05 on Monday 5 October and Manchester’s new Jisc offices are already humming with activity. With coffee in hand, I peruse the weekend’s e-mails and focus on several submissions to the JUSP (Journal Usage Statistics Portal) helpdesk. More than five years into my diverse role on the service, I’m familiar with pretty much any query that could come in from the 180+ institutions and 70+ publishers that are part of the service1, and today’s issues are fairly straightforward.
The first problem is a typical one and concerns ‘missing’ data. JUSP currently holds JR1, JR1a and JR1 GOA (gold open access) usage data for subscribing sites as far back as January 2009 (where available). As new institutions join the service, we ask them to supply us with a list of the JUSP publishers from which they take journal titles; we then work with the institution and/or each publisher to obtain credentials to access the usage stats. In this case the institution didn’t notify us that they had a subscription for the publisher. Having asked the institution for their account information, the data are collected, processed and loaded for them within minutes.
Our second issue of the day looks more complex. A major UK publisher has restated its entire journal usage data for July 2015, meaning that the figures held within JUSP need to be completely replaced. This comprises 420 individual files for 140 institutions, representing some 120,000 statistics in all. By using the SUSHI protocol, I am able to download, process and replace all the affected data for all sites within half an hour! This is one of the real selling points of JUSP: how many of the individual institutions would have been aware of the issue, and, perhaps more importantly, how many would have had the time or staff resource to download the updated data?
The sheer volume of statistics we collect each month means that manually downloading files one at a time or using a publisher’s web interface is simply not possible, so SUSHI compliance is a must. In the early days of working on JUSP, we helped inform the development of many SUSHI services for publishers worldwide, and we continue to help both publishers and third party statistics suppliers to achieve COUNTER compliance; for example, we have helped to test various publishers’ SUSHI services as they moved from COUNTER Release 3 to Release 4 in recent years.
By carrying out such a large amount of data collection, we are able to save JUSP participants a huge amount of time and effort in tasks such as downloading and processing data, freeing up time for their staff to look at the usage and apply it to their purchasing decisions, internal analytics and however else they choose to make use of the stats. This is a particularly important part of the year as the annual SCONUL return is due, so having the most accurate figures is vital – by making use of our SCONUL return report, sites can view and export up to 70 sets of academic year usage figures in seconds.
With a replenished coffee in hand, it’s time to think about my JUSP colleagues and check in to see what they’re doing. JUSP has a diverse team based at different universities and companies across various locations. Service manager Ross MacIntyre, a fellow Manchester-based Jisc colleague, is en route to Oslo to speak at the 2015 ICOLC meeting. He’ll be promoting the JUSP service and its benefits to an international audience. Immediately prior to Ross’s presentation there is one from Britt-Marie Wideberg on behalf of the Swedish BIBSAM consortium.
BIBSAM is the first overseas group for which JUSP is providing a tailored service. Launched in early October 2015, the service provides a very similar implementation of JUSP to the UK version, but with some features and functionality specifically tailored for the Swedish users. This is the first tangible evidence of the type of support we provide to the international usage statistics community – we are happy to work with groups and individual libraries at an advisory level, though BIBSAM is the first group to take advantage of a JUSP-hosted service. We were also actively involved in helping the French COUPERIN statistics service get off the ground by providing guidance on software development and advice on working with publishers and libraries.
A growing feature of working on JUSP is the role we’re starting to play in informing developments across the wider usage statistics and analytics sector. Project manager Jo Lambert is typically at the forefront of this work, and today she’s attending a range of meetings which will cover issues such as JUSP’s continued work to share data and information with our sister service, KB+, the usage of analytics across Jisc as a whole, and ensuring that the work we do on JUSP mirrors related activity in the sector. Jo is also a regular speaker at conferences, both at home and overseas, and has helped promote the service to a huge range of international audiences. Today she’s also following up a recent approach from the ICOLC consortium, which is looking to the JUSP team to provide support and guidance as it works with institutions in the United States to rationalize their own usage statistics provision.
Back in Manchester, my colleague Dave Chaplin and I are starting to look at September data for inclusion in JUSP. Under the COUNTER guidelines, publishers have up to 28 days from the end of the previous reporting period to make the usage data available. However, different publishers release data at different times of the month, depending on their preferred platform or stats provider and various other factors. Typically, we see around 30 new sets of data within the first week of each month, with the other 40 sets coming in at intervals up to the end of the deadline (and very occasionally some slippage if a publisher has any technical issues). On this occasion, my tests of each SUSHI service have confirmed that I can now gather the September data for 26 different publishers – for some publishers we will need to gather data for around a dozen institutions, in other cases there will be 140 or more UK HEIs for which we need to download data. Dave has performed a similar test on the subset of publishers in the BIBSAM implementation, for whom eight new data sets are available.
Having started the major task of gathering some 6,500 files for JUSP and more than 1,000 for BIBSAM, it’s time for a chat with JUSP’s other ‘technical’ team member, Paul Needham. Based at Cranfield University, Paul has been responsible for developing all our SUSHI clients, our in-house SUSHI server, and indeed much of the JUSP portal functionality during its first few years of service. Split across many projects and services, Paul’s time on JUSP is at a premium, and today I need to catch up with him to find out how work is progressing on our development of an e-books companion service to JUSP.
Set to launch in February 2016, JUSP’s e-books service will initially comprise a small set of publishers and (hopefully) e-book aggregators. We will be offering a series of reports and features along similar lines to JUSP, allowing users to not only view their BR1, BR2 and BR3 reports but also see the data manipulated in various ways. My discussion with Paul focuses on some issues we’ve noted with formatting of files from publishers and some issues with various vendors’ SUSHI services. As a member of the COUNTER Technical Advisory Group, Paul is working directly with each vendor to ensure that their reports meet COUNTER Release 4 guidelines – a process we also undertake for JUSP as new publishers join. Today he has spoken to six publishers to advise them of changes required to their files before they will pass all the necessary checks required for loading into the new service. A service like JUSP, with its primary reliance on automated checking functionality, is very much based around standards and compliance with protocols and existing guidelines; it’s therefore vital that we work with vendors to ensure that these are correctly used.
The first seven September data sets have now downloaded back in Manchester. Every file on JUSP is put through a rigorous series of checks, both visual and automated, and they must pass exacting criteria before they can be loaded into the database. This can either be a very swift or a very time-consuming process, depending on any errors found during the checking routines, but today I’m fortunate that much of the initial file processing goes smoothly, with no errors thrown up in the first three data sets. We check for many things, such as correctly formatted titles and identifiers, the presence of totals, correct date formats, a sanity check to ensure that files actually contain real data, and many other tests.
Files that pass these checks are converted into a format ready to load into the JUSP database; once loaded, the data will join the 250 million-plus individual statistics already in the portal. As you can imagine, it’s impossible to manually check all files after loading, and while we run random spot checks on files to see if they have loaded, we also put some of the power of testing data back in the hands of individual libraries. Fortunately, the JUSP community is both extremely proactive, and also communicative; sites know to use the JUSP helpdesk as soon as they spot any issues with their data, and we are very proud of our close links both with institutions and publishers, which enable us to resolve any issues with the minimum fuss or delay. Indeed, the power of a community is important here – should there be a systematic issue with publisher data, the vendors tend to work with us to solve issues quickly as we’re reporting problems on behalf of over 150 sites.
Liaison with sites is very much the remit of another arm of the JUSP team – our colleagues at Evidence Base, located at Birmingham City University. Jo Alcock and Pete Dalton have responsibility for liaising with sites that wish to join JUSP, surveying users periodically on aspects such as publisher ‘wish-lists’, new functionality, developments they’d like to see and how institutions are using the service to inform their collection management. Jo in particular is actively involved in promoting JUSP through the regular, well-attended webinars that we host. We’re currently in the planning phase for 2016, and Jo’s working with various groups, including our CAG (Community Advisory Group) on forthcoming training and support needs. Today she’s also been in touch with an HE college that wants to join JUSP to find out which publishers they subscribe to, and their requirements from JUSP.
We very much empower libraries with the development of JUSP; most of the new reports in the portal have come from community feedback, via the CAG, or from suggestions received during our regular webinars and training courses. If we believe that a change or new addition can benefit the community as a whole, we’ll work to put that into the live service – that’s part of my role. In the case of a new report, or major changes to existing functionality, we will also provide accompanying support materials and guides on the JUSP website, as well as publicizing developments through various mailing lists, on the website and using our Twitter feed @JUSPSTATS.
As lunchtime approaches, I’m now onto the sixth of the new data sets, and this one has thrown up two interesting issues. In order for us to assign identifiers to each item of usage, we have controlled lists of journals to check against, and five titles not currently on the master lists have been flagged up when processing this latest data set. Fortunately, I can simply add these new titles into the database, and run the checking script again; this issue is now fixed. The second problem is slightly trickier to fix. One file has empty print and online ISSN identifiers for a journal entry. As it’s not a systematic error affecting the files of all institutions for the publisher in question, I can fix this by manually editing the source file itself. Normally we prefer publishers to make such changes, but as the issue is affecting just one file from the 450 I’ve downloaded, in this case I’ll edit it and move on. Having put in the identifiers from our master list, the file then passes its checks and can be loaded into JUSP.
These sorts of issues tend to arise when a publisher first joins JUSP. That area is in the hands of our Oxford- and London-based colleagues at Jisc Collections. With the Frankfurt Book Fair fast approaching at time of writing, Vicky Legge is hard at work putting together appointments and meetings during which she’ll extol the benefits of JUSP to e-book and journal vendors, and also meet with existing JUSP-subscribing publishers to check they’re happy with the service provided. For most publishers, working with JUSP has been extremely straightforward. Aside from an initial set-up phase, which usually involves signing a free publisher participation agreement and then providing us with access to the SUSHI service, there are minimal ongoing resource implications for the publishers. We occasionally need to go back to them for updated credentials or to fix any systematic issues, but generally, once a publisher has agreed to join JUSP, things proceed very smoothly. Our users have a constantly updated publisher wish-list, and the Jisc Collections team works with these publishers to bring them into the service. The more publishers that are part of JUSP, the more time and effort we can save libraries and institutions in terms of gathering data and providing annual summaries of their usage.
By 16:00 Dave’s finished loading the BIBSAM data and I’ve also worked through around ten of the September data sets, with the remainder continuing to collect; processing the rest of these will be a job for tomorrow. The final part of my day will be spent on typical monthly admin tasks such as collating September’s usage of the JUSP portal itself. My colleague Hilary Jones also puts together various monthly reports which describe how people have been accessing and using JUSP. This is an extremely valuable piece of work which can help inform aspects of the JUSP service, such as the promotion of reports that do not appear to have gained much usage, and we can easily see if any sites aren’t using the service. We can then put out news items or tweets to promote these reports or other features, or discuss why sites might not be making use of the service we offer.
As 17:00 arrives, Ross has confirmed that he’s in Norway and ready for the following day’s conference, Paul Needham has received responses from several publishers promising to look at the feedback on their e-book report COUNTER compliance, and Jo Lambert has finished her meetings.
With more data collecting overnight, I can log off safe in the knowledge that JUSP continues to serve a demonstrated need for our users, while supporting standards and promoting good practice on an international level. Our adoption of SUSHI for data gathering, combined with rigorous processing checks, ensures that we present data to users in the most accurate and timely way possible, freeing up their time to do something a little more interesting than a mass download of spreadsheets each month!
Abbreviations and Acronyms
A list of the abbreviations and acronyms used in this and other Insights articles can be accessed here – click on the URL below and then select the ‘Abbreviations and Acronyms’ link at the top of the page it directs you to: http://www.uksg.org/publications#aa