This weekend Musicmetric will be at the Silicon Milkroundabout job fair. We were there last year and met some fantastic people. This year we are hoping for more of the same.

So, if you are a talented engineer, designer, or project manager we would love to meet you. If you are none of the above, but take pride in what you do and are looking to join a start up that is making waves in the entertainment world we would love to meet you as well. We will be there all day Saturday and Sunday so come by our stand and say hello. If you can’t make it but are interested in working with us please check out our jobs page.

From Thursday to Saturday Musicmetric will be attending Liverpool Sound City, the largest music, digital and film festival in the UK. As the birthplace of The Beatles, Liverpool has a strong musical history so we are excited to be attending all the great conferences, parties and expos taking place at the festival.

If you’d like to meet up and get a demo of Musicmetric or discuss how metrics can help you, drop us an email at soundcity@musicmetric.com.

We look forward to seeing you there!

Without doubt the problem of “Fake Fans” or “Fake Followers” – social media activity caused by artificially created and controlled social media user profiles – is prevalent and on the rise. Since the Musicmetric platform was built specifically for collecting and processing data from social media, anomaly detection and data cleaning are core parts of our data processing system.

Time series data that is collected by our infrastructure hour by hour is processed by a data cleaning algorithm before being presented to the user. At its simplest, it performs interpolation to accommodate irregularities in the update frequency of a site being monitored. If a site updates the reported cumulative total for fans after a three hour gap, the fan additions will be spread over the three hour period for hourly data.

Additionally, our anomaly detection algorithm analyses a window of time series data in order to detect if a “spike” has occurred in “daily diff” time series. Since we sample data hourly if there is a one-off anomalous event we are able to remove that single time point from the time series and still have high quality data at the daily level. The algorithm can also deal with multiple concurrent spikes for a short period. Where a site is manipulated through the addition of fake likes/followers the algorithm will detect the characteristic shape of this activity and automatically discount the number of followers. On more than one occasion we have received confirmation that our algorithms have successfully corrected for fraudulent activity in this way.

Aside from the measures we take in our system to correct for the actions of “spam” accounts on social media sites, the sites themselves dedicate significant effort to detect and remove these accounts which can result in further anomalies in our time series data as they get deleted. In particularly bad cases, if an artist has a high enough number of spam fans, their fan count may actually decrease, showing a negative change day on day as the spam accounts are steadily removed by the social network.

In addition to the measures already being used in the system, we have recently been performing further data science over all the time series data we have for all the artists we have been tracking over the past four years. By doing so, we can further characterize what constitutes unusual activity in terms of the values in a single time series data set for an artist, but also across different time series for the same artist. It is natural that depending on the artist there may be different activity on different sites, but by looking at all the artists we are tracking across all the sites we can isolate the most unusual activity. With one set of analysis we found that we had a reliable means to detect the event of the death of an artist; truly an event that elicits a non-regular response in social media.

As such, whilst we have many measures in place to tackle manipulation of social media statistics, our efforts are being refined continually, with the results of these efforts feeding into our products all the time.

This year’s Music Hack Day Paris was hosted in Deezer’s stylish new offices in central Paris. After coffee and pastries we were greeted with intros by the organisers Axel Calandre (Deezer) and Martyn Davies (Music Hack Day). Deezer CEO Axel Dauchez then gave an inspirational talk about the importance of context and it’s scope to increase the emotional connection between listeners and both music and musicians, a problem only 5% solved by his reckoning.

There were a brisk set of API presentations from Musicmetric and the other event sponsors followed by hotdogs in the Deezer office garden to fuel us up for the coding to come. The official coding kicked off at 2pm and as is fitting with Music Hack Day tradition, the hack venue was open round the clock to enable a full twenty-four hours of uninterrupted hacking.

The payoff from all the hard work and lost sleep was a veritable fiesta of music related hacks, some useful, some fun, mostly a combination of both. “Giraffe” was the Musicmetric prize winner (of an eggbot) for graph-based analysis of the relationships between artists using Neo4J. Our API was used as a source of semantically annotated articles about musical artists and although the project wasn’t fully completed, we were very impressed with the scale of the project attempted for a twenty-four hour hack.

For our hack we were inspired by the sheer volume of Daft Punk we were hearing at the event so we set out to do some Data Science to explore the popularity of French artists, in France and globally. We started by using Musicbrainz to assemble a list of French artists and then used the Musicmetric API to fetch Bittorrent activity data for downloads both globally and in France in the last week. Using the list of French artists that are being most downloaded globally, we re-sorted the top artists by the proportion of their downloads in France. The aim of this was to find currently popular French artists that were popular with an international audience, but also especially popular in their country of origin – ideal for visitors to the Music Hack Day event.

All in all it was another fantastic Music Hack Day; Deezer were excellent hosts and the organisers did a great job to ensure it all ran smoothly. A complete list of the hacks is available on Hacker League. We can hardly wait until the next one in Barcelona in June, and hope to see you there.

At Musicmetric we love data; it’s the foundation of what we do. We especially like Big Data and in the near future there are lots of data-related events afoot.

Next week is the second annual Big Data Week, a global festival of data with events happening in cities across the world. We are especially active as part of the London data community and co-organise the Big Data London Meetup with our friends at Acunu which kicks the week off so we hope to see you there.

On a more music-related note, this weekend Music Hack Day takes place in Paris hosted at the Deezer HQ. We’re one of the sponsors and will be presenting the Musicmetric API, as well as partaking in some hacking ourselves so we look forward to chatting about music data there. We’ve been involved with Music Hack Days since the first one in London in 2009; if you’re interested in Music and APIs then you should definitely come along. In June there will be the annual Music Hack Day Barcelona in conjunction with the Sónar Festival so there’s still time to prepare for that.

Lastly, following the successful inaugural Strata EU Conference in London last year, this year’s event has been recently announced and the call for proposals is now open. We’re on the programme committee, so if you have an idea for a Big Data related talk get your proposal in by the deadline of 28th May and we look forward to seeing your submission.

South Korean rapper Psy has seen a massive leap in his social media following after the video for his new track “Gentleman” broke the YouTube single day views record by amassing nearly 20 million views in 24 hours, more than double that of the previous record holder Justin Bieber.

On the day the single was released on April 12th Psy gained nearly 23,000 new fans but by the time the video hit YouTube on April 13th he had seen this number grow to over 121,000 new fans per day a staggering 426% increase in daily activity. As of April 14th the rapper had added a further 30,000 new fans and that figure looks set to grow as his YouTube views hit nearly 60 million two days after being uploaded.

The single is the follow up to his monster hit “Gangnam Style” which still holds the record for being the most watched YouTube video of all time with 1.5 billion views.

Check back with Musicmetric as we track Psy’s data to see if “Gentleman” will be able to replicate the success of “Gangnam Style”.

SXSW has built a reputation and an increasingly large community of artists and delegates returning to Austin year after year to be a part of one of the biggest international festivals celebrating new as well as popular music. It is also a conference that celebrates the latest technology as well as film, so Musicmetric saw an opportunity to really inspire and engage with this passionate and creative community. This year, alongside producing our usual SXSW Artist infographic highlighting the top tips amongst this year’s line-up, we also created a real time “Who’s Trending” app in partnership with Billboard.

The app tracked the @mentions happening on Twitter for every artist playing the festival second by second. This meant that we knew exactly which artists were generating the most buzz online pre, post and even during their sets. We knew when gig rumours broke, who really engaged their audience and who disappointed. Our realtime app delivered insight and created a totally new kind of on-the-go filter for all those navigating the maze of venues and artists for seven days.

Now the dust has finally settled in Austin, Texas following last month’s SXSW we have joined forces with Billboard again to take a look at the artists who have seen the biggest surge in popularity during their performances during the music and technology festival. The analysis will uncover insights into how engaged the audiences at SXSW are overall and therefore what benchmarks might be expected of a really engaged audience at a show. It will also reveal which artists managed to sustain growth overall and who had buzz that simply came and went.

Perhaps not a great surprise then that comeback kid Justin Timberlake saw the biggest increase in new fans, gaining 652,000 over the festival – more than twice as many as second place Usher. Timberlake also saw equally significant growth in plays, with over 11 million during the festival. In the run up to his surprise performance at the Myspace Secret Show he experienced two huge spikes in Twitter mentions from 7pm onwards as the rumour mill went into overdrive with attendees speculating about his appearance.

Hip Hop was clearly growing in popularity this year with rap veteran Snoop Lion (previously Snoop Dogg) making regular appearances on our “Who’s Trending” app alongside relative newcomer Trinidad James. The rappers have seen increases of 15% and 19% in total plays respectively. Hot tip Macklemore and newcomer Kendrick Lamar have also seen their fan bases and Twitter mentions grow considerably, with Macklemore gaining 177,000 fans and Lamar gaining 110,000 over SXSW 2013.

YouTube star Tori Kelly, who was famously rejected by Simon Cowell on season nine of American Idol in 2010, saw an increase in daily Myspace views to nearly 4,000 on March 17th, almost twice as many as three days earlier.

As usual SXSW played host to the biggest emerging talent in music such as ZZ Ward and Waxahatchee with rapper Angel Haze gaining a huge 400% increase in new fans in the days following the festival.

Take a look at our infographic for Billboard and get in touch with us if you are interested in more indepth data on this or any other music or artist related event or campaign.

After the dust settled on this year’s Canadian Music Week we thought it fitting that our chairman Jeremy Silver gave us the key points from his speech at the Social Music Summit.

In his opening remarks Jeremy spoke about the growing relationship between music and social media and stressed that record labels need to be more focused with fanshare and less obsessed with market share.

He discussed the continued importance of charts and how they have evolved beyond being based solely on sales to now include YouTube streams, such as the offering from Billboard and asked attendees to consider the significance of Soundcloud and Bittorrent rankings provided by Musicmetric.

He finished with some final points for the attendees asking for a change in music marketing where fans were not controlled but were encouraged to engage in conversations with their favourite artists.

A recent study by the European Commission suggesting piracy has little impact on music sales has been criticised as being “flawed” and “misleading” by industry insiders. The study of 16,000 Europeans was analysed over the course of a year and concludes that “although there is trespassing of private property rights (copyrights), there is unlikely to be much harm done on digital music revenues.”

Musicians and music industry bodies alike have strongly disagreed with the EU findings and the International Federation of the Phonographic Industry (IFPI) have said the report is “disconnected from commercial reality.”

They also argued that the study is “contradicted by a large volume of alternative third party research that confirms the negative impact of piracy on the legitimate music business.”

Musicmetric have referenced IFPI data before in the detailed Digital Music Index and although Musicmetric doesn’t condone illegal downloads, understanding piracy data does help us to determine the demand for an artist and the global geographic spread of fans very accurately.

The IFPI also claim that piracy affects the ability of legal services such as Deezer and Spotify to grow and monetise their audience although research by Musicmetric and others indicates that piracy levels are reduced in countries where legal music services like Spotify are available.

Following the release of his video for chart-topper “Mirrors” Justin Timberlake is currently riding high on the iTunes global album charts with his much lauded effort “The 20/20 Experience”.

After taking a seven year hiatus from music the former teen heartthrob saw a huge spike in his online popularity in January following the release of the lead track from the album “Suit & Tie”.

Ahead of the song’s release Timberlake was adding an average of 200k new plays a day but a week after the release this figure peaked at just under 2 million.

Even more significant was the impact of his Grammy’s performance on statistics. The day after the star studded ceremony on February 10th, Timberlake saw an almost 100% increase in new fans.