The example appears to be a composite image of two images of the Palace of Westminster from the walkway under Westminster Bridge. Both from the same location probably from a tripod and only one including the "firework effect" which is commonly achieved by setting steel wool alight at the end of a rope or line that a subject whirls around during a long exposure.

We can see a few pieces of information in the image which back this up...

• There's no shadow from the figure obscuring Elizabeth Tower (formerly but incorrectly known as St. Stephen's Tower or Big Ben which is actually the Largest Bell inside.) Given the small radius of the spark trails, the subject could not have moved enough to avoid darkening the tower or the clock face.
• The subject wielding the apparatus appears to lack a torso yet their feet clearly remained quite still as they're relatively clear.
• The edit is relatively obvious and has only had a basic feathered blend applied in a circle. This is also visible in sparks which should be bouncing off the roof into the central circle and are instead cut off quite abruptly (and you can seen them all falling across the path of the circular trail.) Had the image been one exposure the houses of parliament would be obscured in a curtain of sparks.
• The image has 2 points of focus as it is sharp in the brickwork, loses sharpness somewhat along the bridge and regains it at the building opposite.

In terms of what you need to create a similar shot, you need a camera which can be attached to a tripod, allow manual control of exposure and allow a multi-second exposure (around 10-15 seconds and above.) Typically that would require a DSLR or EVIL (mirrorless) camera rather than a phone or point & shoot camera. If you search on the terms steel wool long exposure you'll come across a number of tutorials like this one.