This is the first of four innovative tools we're looking at in the Better Twitter Analytics series, seeking a new way of looking at data to help you shape and improve your nonprofit's social media strategy.
Twitter StreamGraphs
is a Java-based visualization tool that can give you a quick overview
of the last 1000 tweets around any search term on Twitter — and that
includes usernames and Twitter lists.
Put in your search term, your username with @ in front of it, or
@username/list and what you get back, assuming that you have Java
enabled on your browser, is a colorful visualization of the tweets
related to your search.
Each of the curving areas in different colors
represents a different related word. Click on any one of those to see
specific tweets listed at the bottom of the page. Along the bottom of
the graph you’ll see the dates, and you can click on any of those to
change which day’s tweets you’re looking at.
So, basically, Twitter StreamGraphs is a search tool that gives you
another way to visualize the results you get when you search for any
keyword on Twitter.
Pretty, yes, but is it useful?
When I first came
across Twitter StreamGraphs a couple weeks ago, I have to admit that I
dismissed it as an interesting toy. Maybe it was the color choice —
pinks and purples — which I found hard to take seriously! But when an
analytics guru finds a tool interesting, you know it probably warrants
a closer look. So, we’re taking Avinash Kaushik’s practical approach to
social media analytics here, where you want to ask yourself two simple questions:
- What is this saying?
- What action can I take?
For an example, let’s take Toastmasters International.
It’s one of those surprisingly rare cases where the Twitter username
and the name by which most people know the organization are the same,
which simplifies things a whole lot!
First, have a look at the StreamGraph for tweets coming from the
Twitter username @Toastmaster, to get an overview of the messages put
out by the organization on Twitter:
It looks like @Toastmasters has been running a trivia contest to
highlight the 85th anniversary of the organization, to judge by the
dominant keywords:
- Anniversary
- Trivia
- Question
- Smedley
- 85th
Let’s set aside the obvious contest-related words — it’s probably
safe to assume this is promotion of a one-time event, not typical of
the account’s tweeting patterns — and look at the rest of the keywords
that show up most prominantly:
- Toastmasters
- Tips
- Job
- Speech
- Club
- Meeting
What do the keywords here convey about the organization? How well
does the content of the tweets correspond with the messages the
organization puts out on its website and through other channels? In
short, do the tweets reflect the brand? If not, there's an action for you!
Now, what are other people saying about the organization on Twitter?
That’s where StreamGraph number two comes in. Here’s what we get for a simple search on “Toastmasters” as a keyword:
Some of the most obvious keywords that jump out from this graph are:
- Toastmasters
- Club
- Speech
- Meeting
- Public
- Speaking
- Training
So, how does this content overview differ from the message that @Toastmasters is
putting out, and where do the two sets of terms overlap?
Click on a few
dates on the StreamGraph to check out the corresponding tweets… What
does this content tell us about the primary interests and concerns of
@Toastmasters’ target audience?
Is there an untapped area of interest that could/should be
addressed? If so, it might be worthwhile for the organization to do a
bit of experimenting — changing the wording of tweets to align more
closely with the keywords used by the audience, and keeping in mind the
interest of the audience in deciding what content to tweet.
No organization is associated with one simple keyword, of course —
not even when that keyword is its name. For a more complete analysis of
your own nonprofit’s Twitter presence, you’d want to add at least one
or two more graphs for comparison, scouting for keywords related to
your mission and clicking on individual words in the StreamGraphs to
find related terms people are using around that topic and what they’re
actually saying.
But the words are only one part of the visualization. Take a look at the two graphs together:
Remember, each graph covers just the last 1000 tweets. The
@Toastmasters account covers a date range of July 13th to December
11th (the day I grabbed this screenshot), while the second graph,
showing the last 1000 public tweets using the word “Toastmasters,” only
goes back to December 3rd.
What do the pictures tell us?
Clearly, the volume of tweets about the
organization by far outweighs those coming from the organization. That
is to be expected, however: After all, it’s one voice compared to many voices.
It would be nice to be able to compare date for date, to see if there's a correlation, but the volume of tweets is too different between the two search terms, so the resulting graphs are way out of scale. That's something to keep in mind for your own nonprofit, however, if your tweets and those related to your keyword search are somewhat more equal, which is more likely to be the case with a smaller organization, with a smaller membership/audience, and with a more targeted keyword search.
A little clicking around reveals that many of the general tweets including the keyword "Toastmasters" are location-specific. This is not surprising, perhaps, as an international
organization tends to interact most directly and most often with its
members at the local club level — that’s often where the
most immediate “relationships” are developed with members, day to day.
To zero in on that local level, we could repeat the Twitter StreamGraph
analysis with user accounts for local clubs or chapters, to
- look at what the
local clubs are doing on Twitter, compared with the parent
organization, and
- compare the club accounts’ content with their own
local audiences.
For example, the Toronto Toastmasters
might generate a StreamGraph for their Twitter username
(@TOToastmasters), and compare that the results for search terms such
as “Toronto Toastmasters” and “Toronto public speaking.”
Again, we’d want to look at the dominant keywords, and how well the
outgoing messages match with what is of most interest to the audience.
What else might you do with this tool?
If you are using a unique hashtag to track a campaign, you could
generate a StreamGraph for your hashtag (without the # sign — this tool doesn’t seem to be able to handle the special character) and see if there’s a relationship
between the frequency of “outside” mentions of that hashtag with the
pattern of your organization’s tweets.
Looking at the audience use of keywords and the specific tweets,
too, you might be able to gain a sense of whether they’re getting their
questions answered, which might suggest topics for future tweets.
And again, it might be useful to adapt the club’s Twitter content to
more closely align with keywords that come up in the general
conversation of the target audience. Worth a try?
Another action might be to explore whether audience tweets about the
topic pick up when the organization increases its tweeting frequency.
If so, that’s probably a good clue as to what’s working for you on Twitter and what’s not...
This isn’t the kind of data you can quantify easily and take into
the boardroom, but if you’re the person in charge of your nonprofit’s
Twitter strategy, there are useful insights to be gleaned from a little
quality time with the Twitter StreamGraphs tool. I’ve just skimmed the
surface here, with the focus on comparing the organization’s outgoing
messages with what’s going on in the audience and looking for points of
relevance. We haven't even looked at what a Twitter Lists search might turn up!
But now it’s your turn to play —
How could your organization use Twitter StreamGraphs? What other information
could be extracted, to help you make your nonprofit's Twitter strategy more effective?